የኮምፒዩተር ማረጋገጫ, ማረጋገጫ እና የኮምፒዩተር ማረጋገጫ እንዴት እንደሚቻል

የቅጂ መብት: Enes Altuncu, ea483@kent.ac.uk (University of Kent, UK) Virginia N. L. Franqueira, V.Franqueira@kent.ac.uk (University of Kent, UK) Shujun Li, S.J.Li@kent.ac.uk (University of Kent, UK) የቅጂ መብት: Enes Altuncu, ea483@kent.ac.uk (ኮንቲ ዩኒቨርሲቲ, ዩኬ) Virginia N. L. Franqueira, V.Franqueira@kent.ac.uk (ኮንቲ ዩኒቨርሲቲ, ዩኬ) Shujun Li, S.J.Li@kent.ac.uk (ኮንቲ ዩኒቨርሲቲ, ዩኬ) Abstract አጠቃቀም በአጠቃላይ, ይህ ጽሑፍ በንግሊዝኛ እና ቻይና ውስጥ የፈጠራ ጽሑፍ እና መሳሪያዎች ላይ የተመሠረተ ነው, ይህ ጽሑፍ ደግሞ በ 2020 እና 2021 ውስጥ የተመሠረተ 12 የተመሠረተ የፈጠራ ጽሑፎች መካከል አንድ ሜታ-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር-የተኮር : Deepfake, Survey, Definition, Datasets, Benchmarks, Challenges, Competitions, Standards, Performance Metrics, እና ተጨማሪ ያንብቡ. መለያዎች 1 መውሰድ የ IA እና ማሽን መግዛት ውስጥ የቅርብ ጊዜ ልምድ የቪዲዮ, ፎቶ, እና የኦዲዮ ወዲያውኑ የበለጠ እውነተኛ መሳሪያዎችን ለማምረት ችሎታን ያካትታል. በተለይም የ state-of-the-art deep learning methods የ "deepfakes" ን ለመፍጠር ይቻላል, የ "deepfakes" ወይም የሲንቲክ መሳሪያዎች ማንኛውም ሰው አጠቃቀም ቀላል ሊታወቁ አይችልም. deepfake (የመጀመሪያ በ 2017 መጨረሻ ላይ ተመሳሳይ ነበር) አንድ ተመሳሳይ አዲስ ተጽዕኖ ነው ቢሆንም, በውስጡ ልምድ አስደሳች ነው. በ 2019 እና 2020 Deeptrace ግምገማዎች መሠረት የ deepfake ሁኔታ ላይ ግምገማዎች [ በንግሊዝኛ ላይ የ Deepfake ቪዲዮዎች ቁጥር ከ 7,964 (የሐሴ 2018) ከ 14,678 (የሐሴ 2019) ወደ 85,047 (የሐሴ 2020) ወደ 968% ይጨምራል. 2 በዚህ ጥናት ውስጥ, የ deepfake-related research ecosystem በተመለከተ የተለያዩ ባህሪያትዎች, በ performance metrics እና standards, datasets, challenges, competitions, and benchmarks ጨምሮ, ይመልከቱ. በተጨማሪም, በ performance comparison, key challenges, and recommendations እንደ በተሳካ መንገድ በተሳካ ሁኔታ ከሁሉም ተጨማሪ ባህሪያት ያካትታል 12 የተመሠረተ deepfake-related survey papers አንድ ሜታ-የተመልከት ያቀርባል. በ "Deepfake" ልምድ ላይ የ "Deepfake" ልምድ እና በ "Deepfake" ልምድ ውስጥ በ "Deepfake" ልምድ ውስጥ በ "Deepfake" ልምድ ውስጥ በ "Deepfake" ልምድ ውስጥ በ "Deepfake" ልምድ ውስጥ በ "Deepfake" ልምድ ውስጥ በ "Deepfake" ልምድ ላይ በ "Deepfake" ልምድ ውስጥ በ "Deepfake" ልምድ ውስጥ በ "Deepfake" ልምድ ውስጥ በ "Deepfake" ልምድ ውስጥ በ "Deepfake" ልምድ ውስጥ በ "Deepfake" ልምድ ላይ በ "Deepfake" ልምድ ላይ በ "Deepfake" ልምድ ላይ በ "Deepfake" ልምድ ላይ 1.1 የኮምፒውተር አጠቃቀም የ "Deepfake" መለያው "Deepfake" (የ "የ" "የ" "የ" "የ" "የ" "የ" "የ" "የ" "የ" "የ" "የ" "የ"" "የ" "የ"" "የ" "የ" "የ" "የ" "የ"" "የ" "የ" "የ"" "የ" "የ"" "የ" "የ" "የ" "የ""" "የ" "የ"" "የ"" "የ"" "የ" """" """""" """""" """"""""" """"""" """"""" """"""" """"""" """""""" """""" """" (DL)) እና "fake". ይህ በአጠቃላይ በአጠቃላይ በአጠቃላይ የአጠቃቀም መሳሪያዎች (ፎቶ, ቪዲዮ እና / ወይም ኦዲዮ) ወይም በአጠቃቀም መሳሪያዎች በመጠቀም አዲስ (ስንቲቲቲክ) መሳሪያዎች ለመምረጥ ለመምረጥ ጥቅም ላይ ይጠቀማል. በአጠቃቀም የተመሠረተ deepfake ውሂብ የአጠቃቀም ምስሎች, fake speech forgeries, እና fake videos that combine both fake images and fake speech forgeries. While having "fake" in the word indicates manipulated or synthesized media, there are plenty of benign applications of the deepfake technology, e.g., for entertainment and creative arts. In this regard, another term "deep synthesis" has been proposed as a more neutral-sounding ]. This new term, however, has not been widely adopted. አግኙን 60 በአሁኑ ጊዜ, የ Deepfakes እና የ Non-Deepfakes መካከል ያለውን መስመር በእርግጥ ቀላል አይሆንም. በአሁኑ ጊዜ የ Deepfakes ማረጋገጫ እና የ Deepfakes ለመፍጠር ላይ አንድ ሰው, በአጠቃላይ ሁለት አስፈላጊ ባህሪያት አለብዎት. በመጀመሪያ, የ Deepfakes ማረጋገጫ በጣም ተመሳሳይ መተግበሪያዎች በ DL ቴክኖሎጂዎችን በመጠቀም የተፈጥሮ የ Deepfakes ማረጋገጫ ጋር ተመሳሳይ ነው. የ Advanced Detection Methods ደግሞ የ DL መተግበሪያዎችን ለማሻሻል ለመተግበሪያ ይጀምራል, ነገር ግን እነርሱም በእርግጥ ማወቅ አይፈልግም እንዴት የ target media (deep or not) የተቋቋቋመ ነው. በተመሳሳይ ደረጃ ላይ, የ Deepfakes ማረጋገጫ የ Deepfakes ማረጋገጫ የ Deepfakes-ተኮር ማረጋገጫ መተግበሪያዎችን ለመፍጠር አይሆንም (እያንዳንዱ ምርምር ባለሙያዎች ይህን ማድረግ ይፈልጋሉ), ነገር ግን ማንኛውም (የተኮር ወይም አይሆንም) የፈጠራ መተግበሪያዎችን ለመቆጣጠር ይቻላል, ነገር ግን አብዛ ሁለተኛው, አንድ ሰው በ "Deepfake" ውስጥ "የተፈጥሮ" ስዕል በ "የተፈጥሮ" መጠቀም አይያስፈልግም, ነገር ግን ማንኛውም "የተፈጥሮ" (የተፈጥሮ) ቴክኖሎጂ ነው, ይህም በጣም አስቸጋሪ የፈጠራ መሳሪያዎችን ለመፍጠር ነው. የ Deepfake እንደ "የቴክኒካዊ ፍጹምነት ... እና ቴክኖሎጂዎች አንድ ስሜት" በመጠቀም የኦዲዮ-የቪዲዮ ማምረቻ ነው. እነርሱ ደግሞ ሁለት አዲስ ግምገማዎችን ይሰጣሉ, እና በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" በ "የተግበሪያዎች" 9 አግኝተዋል የሽያጭ የ 2018 Malicious Deep Fake Prohibition Act (የአልአይአይአይአይአይአይአይአይአይ) በ 1041 (b) (2): “ ” በ "የእነዚህ" በ "የእነዚህ" በ "የእነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እነዚህ" በ "እ የ 2019 Deep Fake Accountability Act 2 (የተፈጠራው የፈጠራው የፈጠራው የፈጠራው የፈጠራው የፈጠራው) በ 1041(n).(3) “ በ "የእነዚህ" የቪዲዮ ውሂብ, የቪዲዮ ውሂብ, የድምጽ ውሂብ, የኤሌክትሮኒክ ፎቶ, ወይም ፎቶ, ወይም ማንኛውም የቴክኖሎጂ ውሂብ ውሂብ ወይም ውሂብ ውሂብ ነው. (A) which appears to authentically depict any speech or conduct of a person who did not in fact engage in such speech or conduct; and ” (B) የእርስዎን ምርት በከፍተኛ ደረጃ የቴክኒክ መሳሪያዎች ላይ የተመሠረተ ነው, ነገር ግን ሌላ ሰው በጽሑፎች ወይም በጽሑፎች ይህን ሰው ለመፍጠር ችሎታ ነው. በ "Deepfake" ከላይ የተወሰነ የክፍያ ዝርዝሮች ውስጥ ሊሆን ይችላል, DL እንደ ቴክኖሎጂ መጠቀም በዚያም አጠቃቀም አይደለም. እዚህ ላይ በ "እውነተኛነት", "የእውነተኛነት" እና (እያንዳንዱ) "ቴክኒካዊ መሳሪያዎች" ላይ ትኩረት ነው. 1.2 ልኬቶች እና ልኬቶች በከፍተኛ ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት ፍጥነት እና በከፍተኛ ፍጥነት ፍጥነት የ Deepfake በከፍተኛ ደረጃ ላይ ሊሆን ይችላል ቢሆንም, በዚህ ሥራ ውስጥ, ከብዙ ጥያቄዎች ለመጠበቅ ይበልጥ ጠንካራ ትኩረት አለን. ለሜትሪክስ እና መደበኛዎች, በአጠቃላይ የተጠቃሚ የ ML መሳሪያዎች እና የ deepfake ለመፍጠር ወይም ማረጋገጫ መሳሪያዎች ለማረጋገጥ የተመሠረተ መሳሪያዎች ሁሉ ያካትታሉ. በዲፕፋይክ ጋር የተያያዙ ውሂብ ስብስቦች, ግምገማዎች, ክወናዎች እና በባህር ማረጋገጫዎች ጋር የተያያዙ ውሂብ ስብስቦች እና የ "ዲፕፋይክ" ወይም ተለዋዋዋጭ ውሂብ ጋር የተያዙ ውሂብ ስብስቦች ያካትታሉ. ለሜታ-የተግበሪያ, እኛ ብቻ የኮምፒዩተር ጽሑፎች ያውቃሉ, የኮምፒዩተር ጽሑፎች በሜታ ውሂብ (የተግበሪያ መለያዎች, መለያዎች እና የኮምፒዩተር ጽሑፎች) ውስጥ የ "Deepfakes" ምልክት ይመጠራሉ. 2 Methodology በዚህ ግምገማ ውስጥ የተወሰነ ምርምር ጽሑፎች (እነዚህ, የ deepfake-related ግምገማ ጽሑፎች) በ Science Databases, Scopus እና China Online Journals (COJ) ላይ የተመሠረተ ፍለጋዎች በመጠቀም የተመሠረተ ነበር. በ Scopus እና COJ ላይ የሚፈልጉትን ጥያቄዎች ይጠቀማሉ. 3 (diepfake* OR deep-fake* OR “deep fake*”) AND (የተግበሪያ ወይም ግምገማ ወይም ግምገማ ወይም systemati* OR SoK) (የእውነተኛ ምልክት ወይም ምልክት ምልክት ወይም ምልክት ምልክት) የ 41 ግምገማዎች በንግሊዝኛ እና የ 15 ግምገማዎች በቻይና ውስጥ ተመልካች ፡፡ እነዚህ ግምገማዎች ከ 8 ግምገማዎች በንግሊዝኛ እና 4 ግምገማዎች በቻይና ውስጥ ተመልካች ፡፡ የ Deepfake-related Challenges, Competitions and Benchmarks በከፍተኛ ጥቅሞች ጋር የተወሰነ ወረቀት, የኮምፒውተር የግል አጠቃቀሞች ከ ምርምር ወረቀት, የ Google Web searches, እና በ 2020 እና 2021 (እነዚህን Challenges እና Computational Linguistics አጠቃቀሞች በከፍተኛ ጥቅሞች ጋር የተወሰነ ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር , ICCV, CVPR, AAAI, ICML, ICLR, KDD, SIGIR, WWW, እና ሌሎች ብዙ ሰዎች. በተጨማሪም, የተመሠረተ ግምገማ ወረቀት እና የተመሠረተ ፍላጎቶች, የስርዓት, እና ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠረተ ተመሠ 4 3 Deepfake-related Performance Metrics እና መደበኛዎች በዚህ ግምገማው ውስጥ, በ Deepfake Generation እና Detection Methods ላይ አፈጻጸም ግምገማዎችን እና ተለዋዋጭዎችን ያተኮረለን. እነዚህን አፈጻጸም ግምገማዎች የሚጠቀሙትን ሜትሪኮች የእኛን ግምገማዎች ላይ ይገኛሉ. በዚህ ክፍል ውስጥ, በ Deepfake Generation እና Detection Algorithms ላይ የሚጠቀሙትን አፈጻጸም ሜትሪኮች ይመልከቱ. በዚህ ክፍል ውስጥ የሚጠቀሙትን ሁሉንም ሜትሪኮች በተጨማሪም በስፋት በከፍተኛ አፈጻጸም ሜትሪኮች ላይ በጣም ዝቅተኛ የሞትሪኮች እንደ መመልከት ይችላሉ. በመጨረሻው ክፍሎች ውስጥ, እኛ ደግሞ የተወሰነ አፈጻጸም ሜትሪኮች በፈጠራ መደበኛ ደረጃዎች ውስጥ እንዴት ያካትታሉ. በ "ፈጠራ መደበኛ ደረጃዎች" እኛ በፈጠራ ሂደት መሠረት የተመሠረተ መደበኛ ደረጃዎችን ያካትታሉ, አብዛኛውን ጊዜ በ International Organization for Standardization (ISO) እንደ አንድ ወይም ከሁሉም የተመሠረተ መደበኛ ደረጃዎችን ያካትታሉ. በ International Electrotechnical Commission (IEC) እባክዎን በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ ግምገማዎች እና የ ISO Technical Reports (TRs) 5 6 7 4.1 የሙዚቃ Matrix Deepfake detection በዋናነት binary classification problem ነው.A binary classifier takes an input that is a binary classification problem. ወይም እና አንድ binary ዋጋ ያደርጋል ወይም ለምሳሌ, አንድ deepfake ማረጋገጫ ስርዓት የሚፈልጉትን ፎቶን እንደ መውሰድ ሊሆን ይችላል. ወይም ምርት ወይም . በእርግጥ ተስማሚ በእርግጥ ከባድ ተስማሚ ግምገማ አነስተኛ አነስተኛ አነስተኛ በእርግጥ የፈጠራ እውነተኛ ፈቃድ ፈቃድ እውነተኛ ስሜት አንድ binary classifier ለማረጋገጥ ጥቅም ላይ አንድ አስፈላጊ መሣሪያ ነው. ይህ የሙዚቃ ሞዴል ስኬታማነት እና ስኬታማነት ያካትታል. አንድ አሲስ ላይ ሁለት ወዲያውኑ ከሁለት ደረጃዎች እና ከሁለት ደረጃዎች አጠቃቀም: የክፍያ ነው (የእውነተኛ ስኬታማ እና ስኬታማ ስኬታማ) በአሁኑ ጊዜ በአሁኑ ጊዜ እና በአሁኑ ጊዜ ስኬታማ ስኬታማ ስኬታማ ስኬታማ ነው. (የእውነተኛ ውጤታማ እና የእውነተኛ ውጤታማ) በእውነተኛ እና የፈጠራ ውጤታማ መጠን ተመሳሳይ አይደለም. የዲጂታል ዲግሪ (Defacetor) የዲጂታል ዲግሪ (Defacetor) ለ የይነገጽ ማተሚያ ማተሚያ ያቀርባል. የ 2 ሴሎች በዝቅተኛ, TP (የዲጂታል ዲግሪ ማተሚያ) በ TN (አንድ ቁጥር) ), ትክክለኛ ግምገማ ውጤቶችን ያካትታል, እና በድብቅ ውስጥ ሁለት ካርዶች, FN (የ ቁጥር ) እና FP (የ ቁጥር ), ከባድ ምልክት ውጤቶችን ለመፍጠር ጊዜ ሁለት የተለያዩ ዓይነት ስህተት ያካትታል. confusion matrix አጠቃላይ የተመሠረተ ትክክለኛ / ትክክለኛ / እውነተኛ አግኝቷል / አግኝቷል / አግኝቷል 1 true positives true negatives false negatives false positives ምድብ 1: የ Deepfake ማረጋገጫ ለማግኘት binary classifier ለ confusion matrix. fake (predicted) real (predicted) fake (actual) TP FN real (actual) FP TN ትክክለኛ (አሁኑ) የ TP FN እውነተኛ (አሁኑ) የ FP አዲሱ 3.2 ትክክለኛነት እና አስቀምጥ በካቲት ውስጥ የተመሠረተ አራት ዋና ግምገማዎች ላይ የተመሰረተ , ይህም TP, TN, FP እና FN, እኛ አንድ binary classifier ለ ሁለት አስፈላጊ አፈጻጸም ሜትሪክ ያካትታል - እና . 3.1 precision recall አንድ binary classifier ን ትክክለኛነት እንደ ብዛት ነው. ከሁሉም ሻጋታዎች መካከል የፕላስቲክ መሣሪያዎች እና የፕላስቲክ መሣሪያዎች (የፕላስቲክ መሣሪያዎች እና የፕላስቲክ መሣሪያዎች) በፕላስቲክ መሣሪያዎች ውስጥ ሊሆን ይችላል. አግኙን በእርግጥ ተስማሚ ተስማሚ ግምገማዎች 1 የ positive እና negative ናሙናዎች መካከል የ "እውነተኛ" ጥራት የሙከራ ተከታታይ ከከፍተኛ ተለዋዋጭ ነው ጊዜ, ብዙውን ጊዜ የ falses positives ክብደት መተግበሪያ ይጠቀማል, ይህም የሙከራ ተከታታይ ያደርጋል (WP) በ Eq. ( በዚያ 0 በከፍተኛ መጠን እና በከፍተኛ መጠን እና በከፍተኛ መጠን እና በከፍተኛ መጠን እና በከፍተኛ መጠን እና በከፍተኛ መጠን እና በከፍተኛ መጠን ነው. weighted precision 2 አግኙን አንድ binary classifiers አጠቃቀም ነው. በኤምኤምኤም ውስጥ በዩናይትድ ስቴትስ ( በይነገጽ Matrix ውስጥ, ይህ የመጀመሪያው ተከታታይ ውስጥ እውነተኛ ናሙናዎች መጠን ነው. ተስማሚ ግምገማ በእርግጥ ተስማሚ 3 አንድ ምሳሌ binary classifier ይመልከቱ, ይህም የ deepfake እና እውነተኛ (ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወ 3.3 ትክክለኛ እና ትክክለኛ ውጤቶች የፈጠራ ጥቅሞች ላይ ተመሠረተ, እኛ ደግሞ ሁለት ሜትሪኮች መውሰድ ይችላሉ: (TPR) በተጨማሪም ይደውሉ (CDR), as the fraction of the predicted positive samples among the actually positive samples and (FPR) በተጨማሪም ይደውሉ (በአንድ ጊዜ ከባድ ከባድ ከባድ ከባድ ከባድ ከባድ ከባድ ከባድ ከባድ ከባድ ከባድ ከባድ) ) እና true positive rate correct detection rate false positive rate false alarm rate 4 ( በይነገጽ Matrix ውስጥ, TPR የመጀመሪያው ተከታታይ ውስጥ የተመሠረተ ተስማሚ ናሙናዎች ክፍሎች ነው እና FPR ሁለተኛው ተከታታይ ውስጥ የተመሠረተ ተስማሚ ናሙናዎች ክፍሎች ነው. (አንድ ጊዜ) )). 5 recall 3 3.4 እውነተኛ እና በባህር አነስተኛ ዋጋዎች ትክክለኛ እና ትክክለኛ ትክክለኛ ደረጃዎች ጋር ተመሳሳይ, ከባድ የተመሠረተ ውጤቶች ላይ የተመሠረተ ሁለት ሌሎች ትክክለኛ ደረጃዎችን ማረጋገጥ ይችላሉ: (TNR) በአሁኑ ጊዜ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ (ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም.ኤም ) and ( አግኙን true negative rate false negative rate 6 7 3.5 ጥንካሬ እና ልዩነት በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ (ኤስኤምኤስ) እና (TNR) እነዚህ ሁለት ግምገማዎች መካከል ትክክለኛነት ሁለት ዓይነት ላይ ነው. እነዚህ በከፍተኛ-fake-related ምርምር ውስጥ ዝቅተኛ ይጠቀማሉ, ስለዚህ, በዚህ ጽሑፍ የመጨረሻ ውስጥ እኛም እነሱን አይነቶችም. sensitivity specificity 3.6 ተመሳሳይ ፍለጋ መጠን የ FPR እና የ FNR መተግበሪያዎችን ይጠቀማል. እነዚህ ሁለት ደረጃዎች በአብዛኛው ደረጃ ላይ ይቀላቀሉ. ስለዚህ, በአብዛኛው ደረጃ ላይ ይቀላቀሉ, ይህም በአብዛኛው ደረጃ ላይ ይቀላቀሉ, ይህም በአብዛኛው ጊዜ ሊሆን አይችልም, ይህም በአብዛኛው ጊዜ ሊሆን አይችልም, በአጠቃላይ ትክክለኛ መተግበሪያዎች ውስጥ አብዛኛውን ጊዜ ይበልጥ ትክክለኛ መተግበሪያ ነው. አንዳንድ መተግበሪያዎች ውስጥ, ለምሳሌ biometrics, ሰዎች በእያንዳንዱ የሚታወቀው "የመሠረተ" ይወዳሉ. (ወደ) ወይም የ EER / CER አንዳንድ መተግበሪያዎች ለሁለት ዓይነት ስህተት የተለያዩ ደረጃዎች ናቸው, ለምሳሌ, አስፈላጊ deepfakes (እን.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ.ኤ equal error rate crossover error rate 3.7 ትክክለኛነት እና F-Score ከ EER / CER, በተጨማሪም ሁለቱም ዓይነት ፍለጋዎችን ያተኮሩ የሚፈልጉትን ሌሎች ሜትሪኮች አሉ, አንድ binary classifiers አጠቃላይ አፈጻጸም በጣም የተካተተተ ማረጋገጫ ለማቅረብ. እና (እውነተኛ ሁለቱም ሜትሪኮች በአብዛኛው ክፍሎች (TP, TN, FP, እና FN) ላይ የተመሠረተ ሊሆን ይችላል. accuracy F-score F-measure አንድ binary classifiers ትክክለኛነት እንደ ብዛት ነው. አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ አጠቃላይ ( አግኙን ትክክለኛነት የተመሠረተ 8 አንድ binary classifier's F-score በእርግጥ የሜትሪኮች መካከል አንድ ቤተሰብ ነው. የፈጠራ ቅርጽ አንድ ፓራሚተር ላይ የተመሠረተ ሊሆን ይችላል. በዩናይትድ ስቴትስ ( አግኙን β 9 ሁሉንም F-ስኬቶች ውስጥ በጣም የተጠቃሚ ስሪት ነው. በ F-Score ጋር የተሰራ ነው = 1. ተጨማሪ ትክክለኛነት, ይህ በ Eq. ( አግኙን F1-score β 10 3.8 አጠቃቀም አጠቃቀም አጠቃቀም አጠቃቀም አጠቃቀም አጠቃቀም አጠቃቀም (ROC) curves በአብዛኛው ጊዜ ምናልባት አንድ ስኬት (ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ-ወደ Receiver operating characteristic የሚከተሉትን ይመልከቱ. be the set of all test samples and let the output scores ( ) (for all ∈ (በአውሮፕላስ ውስጥ በይነገጽ) ] on the real line. Let ∈ [ ይህ ሞዴል ለወደመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመመ ∈ : S f s s S a, b t a, b s S It is easy to see that, for = , all the samples will be classified as positive, leading to FN = TN = 0 so TPR = FPR = 1; while for = አግኙን , all the samples will be classified as negative, leading to FP = TP = 0 so TPR = FPR = 0. For other threshold values between and , the values of TPR and FPR will normally be between 0 and 1. By changing ከ to continuously, we can normally get a continuous curve that describes how the TPR and FPR values change from (0,0) to (1,1) on the 2D plane. This curve is the ROC curve of the binary classifier. t a t b a b t a b For a random classifier, assuming that ( ) distributes uniformly on [ ] for the test set, we can mathematically derive its ROC curve being the TPR = FPR line, whose area under the ROC curve (AUC) is 0.5. For a binary classifier that performs better than a random predictor, we can also mathematically prove that its AUC is always higher than 0.5, with 1 being the best possible value. Note that no binary classifier can have an AUC below 0.5, since one can simply flip the prediction result to get a better predictor with an AUC of 1 − AUC. The relationship between the ROC and the AUC is graphically illustrated in Figure . f s a, b 1 3.9 ግምገማዎች Another widely used performance metric for binary classifiers that can return a probability score for the predicted label is . For a binary classification with a true label ∈ {0*,* 1} እና የፈጠራ ምናልባት = Pr( = 1), the log loss per sample is the negative log-likelihood of the classifier given the true label, defined as shown in Eq. ( ). log loss y p y 12 Given a testing set with samples, the log loss score of a binary classifier can be calculated using Eq. ( በዚያ is 1 if the -th sample is true and 0 if false, and ˆ is the predicted probability of = 1 ነው. n 13 yi i y i yi 3.10 Extension to Multi-class Classifiers All metrics that are defined based on the four basic values TP, TN, FP and FN can be easily extended to by considering the prediction to be true or false individually with respect to each class. For example, if the system is classifying animals (cats, dogs, horses, lions, tigers, etc.), then a true positive prediction of an image to be of a cat, would simultaneously be true negative predictions for the remaining classes (dogs, horses, lions, tigers, etc.). If an image of a cat is incorrectly predicted to be that of a dog, it would be a false negative with respect to a cat, a false positive with respect to a dog, and a true negative with respect to all other classes. multi-class classification 3.11 Perceptual Quality Assessment (PQA) Metrics By definition, the main goal of deepfakes is to make it hard or impossible for human consumers (listeners or viewers) to distinguish fake media from real media. Therefore, when evaluating the quality of deepfake media, the quality perceived by human consumers of the media is key. This calls for subjective assessment of the perceptual quality of the deepfake media as the “gold standard”. The most widely used subjective perceptual quality assessment (PQA) metric for audio-visual signals is (MOS), which has been widely used by the signal processing and multimedia communication communities, including digital TV and other multimedia-related consumer applications. As its name implies, MOS is calculated by averaging the subjective scores given by a number of human judges, normally following a numerical scale between 1 and 5 or between 0 and 100. MOS has been used in some deepfake-related challenges (see Section ) and also for evaluating and comparing the quality (realness/naturalness) of deepfake datasets (see Section ). mean opinion score 5.2 4.6 As a general subjective PQA metric, MOS has been standardised by the ITU . There are also ITU standards defining more specific subjective Video Quality Assessment (VQA) metrics and the standard procedures one should follow to conduct VQA user studies, e.g., ITU-T Recommendation P.910 “Subjective video quality assessment methods for multimedia applications” . Note that the ITU standards focus more on traditional perceptual quality, i.e., how good a signal looks or sounds, even if it looks or sounds not real (e.g., too smooth). On the other hand, for deepfakes, the focus is rather different because what matters is the realness and naturalness of the created media, i.e., how real and natural it looks or sounds, even if it is of low quality. To some extent, we can also consider realness and naturalness as a special aspect of perceptual quality. 8 9 One major problem of subjective PQA metrics like MOS is the need to recruit human judges and to have a well-controlled physical testing environment and protocol, which are not easy for many applications. To help reduce the efforts and costs of conducting PQA-related user studies, various objective PQA metrics have been proposed, where the term “objective” refers to the fact that such metrics are human-free, i.e., automatically calculated following a computational algorithm or process. Depending on whether a reference exists, such objective PQA metrics can be largely split into three categories: full-reference (FR) metrics (when the original “perfect-quality” signal is available as the reference), reduced-reference (RR) metrics (when some features of the original “perfect-quality” signal are available as the reference), and no-reference (NR) metrics (when the original signal is unavailable or such an original signal does not exist). For deepfakes, normally NR or RR metrics are more meaningful because the “fake” part of the word means that part of the whole data does not exist in the real world, hence a full reference cannot be obtained. RR metrics are still relevant because deepfakes are often produced for a target’s specific attributes (e.g., face and voice), where the reduced reference will be such attributes. NR metrics will be useful to estimate the realness and naturalness of a deepfake, simulating how a human judge would rate it in a controlled subjective PQA user study. PQA is a very active research area and many PQA metrics have been proposed, some of which have been widely used in real-world products and services, e.g., (MSE), (ኤስ.ኤስ.ኤስ) እና (SSIM) for FR PQA of digitalimages and videos defined as in Eqs. (14), (15), and (16), respectively, where X = {xi} n i is the reference (the original signal), Y = {yi} n i is the signal whose visual quality is assessed, n is the number of pixels in X and Y , L is the maximum possible pixel value of X and Y (e.g., 255 for 8-bit gray-scale images), c1 = (k1L) 2 and c2 = (k2L) 2 ) are two stabilising parameters (k1 = 0.01 and k2 = 0.03 by default). For more about PQA metrics for different types of multimedia signals, we refer readers to some relevant surveys [3, 51, 72]. mean squared error peak signal-to-noise ratio structural similarity index measure 3.12 More about Standards Many of the basic performance metrics described in this section have been widely used by deepfake researchers as de facto standards, e.g., EER, log loss and MOS have been widely used in deepfake-related challenges (see Section ). Also, the combination of precision, recall and F1-score has been widely used to assess performance of binary classifiers. While there have been a number of ITU standards on PQA to date, there does not seem to be many standardisation efforts on the performance metrics for evaluation of binary classifiers. This was the case until at least 2017, when ISO and IEC jointly set up the ISO/IEC JTC 1/SC 42 , a standardisation subcommittee (SC) focusing on AI under ISO/IEC JTC 1 , the joint technical committee for standardising “information technology”. 5 10 11 One recent effort that ISO/IEC JTC 1/SC 42 made is to produce the ISO/IEC TR 24029-1:2021 “Artificial Intelligence (AI) – Assessment of the robustness of neural networks – Part 1: Overview” , a technical report (TR) that systematically covers many commonly used performance assessment concepts, methods and metrics. Although the technical report has “neural networks” in its title, most performance assessment concepts, methods and metrics included are common ones for all supervised machine learning models. 12 In terms of performance metrics, two other ongoing work items of the ISO/IEC JTC 1/SC 42 that deserve attention are as follows: ISO/IEC DTS (የቴክኒካዊ ባህሪያት ፕሮጀክቶች) 4213 “ይነገጽ ቴክኖሎጂ – የቴክኒክ ኢንጂነሪት – የሞክሮ መግዛት ክወና አፈጻጸም ማረጋገጫ”13 ISO/IEC AWI (Approved Work Item) TS (Technical Specifications) 5471 “Artificial intelligence – Quality evaluation guidelines for AI systems” 14 While the ISO/IEC JTC 1/SC 42 was created very recently, another standardisation subcommittee under ISO/IEC JTC1 has a much longer history of nearly 20 years: the ISO/IEC JTC 1/SC 37 that focuses on biometrics-related technology. This standardisation subcommittee is highly relevant for deepfake since deepfake faces can be used to spoof biometrics-based user authentication systems. In this context, the following three standards are of particular relevance: 15 : This standard covers general metrics about evaluating biometric systems. Two major metrics in this context are (FAR) and (FRR), which refer to the standard FPR and FNR, respectively. This standard also deprecates the use of single-number metrics including the EER and AUC (which were widely used in biometrics-related research in the past). ISO/IEC 19795-1:2021 “Information technology – Biometric performance testing and reporting – Part 1: Principles and framework” 16 false accept rate false reject rate : ይህ መደበኛ በአጠቃላይ ቅርጸት ያካትታል (PAD) mechanisms, where the term “ ” refers to the “ ”. It focuses on biometric recognition systems, where a PAD mechanism is a binary classifier trying to predict presentation attacks (also called attack presentations, e.g., fake faces) as positive and bona fide (real) presentations as negative. ISO/IEC 30107-1:2016 “Information technology – Biometric presentation attack detec-tion – Part 1: Framework” 17 presentation attack detection presentation attack presentation of an artefact or of human characteristics to a biometric capture subsystem in a fashion intended to in-terfere with system policy : This standard defines a number of special performance metrics for evaluating PAD mechanisms standardised in the ISO/IEC 30107-1:2016. Three such metrics look at error rates: (APCER) referring to the standard FPR, (NPCER/BPCER) referring to the standard FNR, and (ACER) that is defined as the average of the APCER and the NPCER/BPCER. Such metrics have been used in biometrics-related challenges such as Face Anti-spoofing (Presentation Attack Detection) Challenges . When deepfake images or videos are used to spoof a biometric system, such standardised metrics will become relevant. ISO/IEC 30107-3:2017 “Information technology – Biometric presentation attack detection – Part 3: Testing and reporting” 18 attack presentation classification error rate normal/bona fide presentation classification error rate average classification error rate 19 3.13 Discussion: Performance Metrics & Standards This section provided a comprehensive summary of performance metrics used for evaluating and bench-marking binary classifiers. It is rare that all such metrics are used for a specific application. Instead, one or several are chosen based on specific needs. For a deepfake detection system as a binary classifier, many researchers have chosen to use overall metrics such as accuracy, AUC, EER and log loss, but the combination of precision, recall and F1-score is also common. Some deepfake-related challenges and competitions have introduced their own specific metrics, some of which will be described in Section . The use of different performance metrics can make comparison of different reported results more difficult, so we hope the expected new ISO/IEC standard particularly ISO/IEC 4213 will help. 5 It is worth mentioning that, in addition to evaluating performance of deepfake detectors, the introduced performance metrics for evaluating binary classifiers can also be used to evaluate performance of deepfake generation methods by considering how deepfake detectors fail. For instance, organisers of the Voice Conversion Challenge 2018 and 2020 used this approach to benchmark how well voice conversion (VC) systems can generate high-quality fake speech samples. Another point we would like to mention is that for deepfake videos there are two levels of performance metrics: those at the frame level (metrics of each frame), and those at the video level (metrics for the whole video). Generally speaking, the latter can be obtained by averaging the former for all frames, potentially following an adaptive weighting scheme, so that more important (key) frames will be counted more. 4 Deepfake-Related Datasets In this section, we cover all deepfake-related datasets we identified from the meta-review of deepfake-related survey papers, deepfake-related challenges, competitions and benchmarks covered, one online collection of deepfake-related datasets on GitHub , and the co-authors’ personal collections. Table shows basic information about these datasets. We explain them in four categories: deepfake image datasets, deepfake video datasets, deepfake audio/speech datasets, and hybrid deepfake datasets (mainly mixed image and video datasets). 20 2 Note that many datasets of real (authentic) media were also used by deepfake researchers for two purposes. First, any detectors would need both fake and real media to demonstrate their performance. Second, real media have also been used to train deepfake generators as the training set. In this section, we include only datasets containing deepfake media, some of which contain both deepfake and real media. Some datasets, especially those created for deepfake-related challenges and competitions, have separate subsets for training and evaluation (testing) purposes. The split is necessary for such challenges and competitions, but not very useful for people who just want to use such datasets. Therefore, in this section when introducing such datasets we will ignore that level of details and focus on the total number of data including the number of real and fake samples. 4.1 Deepfake Image Datasets [ ]: This dataset contains 4,310 images, including 2,300 real images and 2,010 fake images created using FaceSwap and the SwapMe iOS app (now discontinued). SwapMe and FaceSwap dataset 78 21 [ ይህ ውሂብ ስብስቦች 131,500 ፎቶዎች ያካትታል, በ FaceForensics ውሂብ ስብስቦች ውስጥ ከ 150 ቪዲዮዎች ከ 78,500 ፎቶዎች እና ከ YouTube ከ ከ 150 ፎቶዎች ከ ከ 53,000 ፎቶዎች ያካትታል. Fake Faces in the Wild (FFW) dataset 32 : This is a number of commercial datasets provided by the Generated Media, Inc., with up to nearly 2.7 million synthetic face images generated by StyleGAN. A free edition with 10,000 128x128 synthetic images is made available for academic research. The website also provides an interactive face generator and an API የ generated.photos ውሂብ ተከታታይ የተሻለ የተለያዩነት አላቸው: አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አነስተኛ አ generated.photos datasets 22 23 24 [ ]: This dataset includes 19,457 face images, including 7,948 deepfake images generated from on 175 forged videos collected online and 11,509 real face images collected from various online sources. (Table 2 of the paper shows the dataset size is 19,509, but the dataset downloaded from pCloud contains just 19,457 images.) MesoNet Deepfake Dataset 1 [ ]: This dataset includes 100,000 synthesised face, bedroom, car and cat images by a GAN generator trained based on real images in the FFHQ and LSUN datasets (three object types – bedrooms, cars and cats – for the latter). Note that the name “100K-Generated-Images” was not a proper one as the authors [ ] just used this to name a sub-folder of their Google Drive shared space, but it was used in one of the survey papers [ አግኙን 100K-Generated-Images 30 25 26 30 65 [ ]: This dataset contains 420,053 images of celebrities, including 156,930 real ones downloaded using Google Image API and 263,123 fake face-swapped ones created using two different methods (Nirkin’s method and Auto-Encoder-GAN) Ding et al.’s swapped face dataset 17 [ ]: This dataset includes 87,000 224x224 face images, generated by processing some StyleGAN-generated synthetic images using the GAN-fingerprint Removal approach (GANprintR) proposed by . It is the replaced version of the በ GANprintR አንድ የቀድሞው ስሪት በመጠቀም የተመሠረተ 150,000 ፎቶዎች ያካትታል. iFakeFaceDB 48 Neves et al. FSRemovalDB [ [ ]: ይህ ውሂብ ስብስቦች 40,000 ፎቶዎችን ያካትታሉ, ከባድ እና ከባድ ፎቶዎችን ያካትታሉ. ውሂብ ስብስቦች የ CelebA-HQ ውሂብ ስብስቦች , the Flickr-Faces-HQ dataset , the 100K-Faces dataset (not available any longer, see the description of generated.photos datasets), and . Faces-HQ 21 27 28 29 thisperson-doesnotexist.com [ ]: This dataset includes 625,537 synthesised face images of 10,177 celebrities, with 43 rich attributes on face, illumination, environment and spoof types. The real images were selected from the CelebA dataset . The 43 attributes include 40 for real images, covering all facial components and accessories (e.g., skin, nose, eyes, eyebrows, lip, hair, hat, eyeglass), and 3 for fake images, covering spoof types, environments and illumination conditions. CelebA-Spoof 75 30 [ ]: This dataset contains 299,039 images, including 58,703 real images sampled from three datasets (FFHQ , CelebA and FaceForensics++ ) and 240,336 fake ones in four main facial manipulation types (identity swap, expression swap, attribute manipulation, and entire synthesis). The images cover two genders (male and female), a wide age groups (the majority between 21 and 50 years old), and both low- and high-quality levels. Diverse Fake Face Dataset (DFFD) 11 31 32 33 4.2 Deepfake Video Datasets አግኙን ]: This dataset contains 620 deepfake face videos, generated by face swapping without manipulation of audio, covering 32 subjects and two quality levels (high and low). DeepfakeTIMIT 35 (FF) [ ]: This dataset contains 1,004 face videos with over 500,000 frames, covering various quality levels and two types of facial manipulation. This dataset is now replaced by the larger FaceForensics++ dataset (see below). FaceForensics 55 (FF++) [ ]: This dataset contains 5,000 face videos with over 1.8 million manipulated frames, including 1,000 real videos (with 509,914 frames) downloaded from YouTube, and 4,000 fake videos created using four face manipulation methods (Deepfakes, Face2Face, FaceSwap and NeuralTextures). The videos cover two genders (male and female), and three quality levels (VGA/480p, HD/720p, and FHD/1080p). FaceForensics++ 56 [ ]: This dataset contains 98 face videos, half (49) are real ones downloaded from Youtube, and the other half are fake ones generated using the FakeApp mobile application (which is now discontinued). The video dataset was created to used to demonstrate a deepfake video detection method based on detection of eye blinking behaviours, so all videos contain at least one eye-blinking event. All fake videos were created by swapping the original face in each of the real videos with the face of the actor Nicolas Cage , thus, only one subject is represented. UADFV dataset 39 34 [ ]: This dataset contains 142 “in the wild” deepfake portrait videos, collected from a range of online sources including news articles, online forums, mobile apps, and research presentations. The videos are diverse, covering the source generative model, resolution, compression, illumination, aspect-ratio, frame rate, motion, pose, cosmetics, occlusion, content, and context. Deep Fakes Dataset 10 [ ]: This dataset contains 5,244 face videos of 66 subjects with both face and voice manipulation. It was released as a preview of the full dataset of the 2020 Deepfake Detection Challenge (DFDC, see below). DFDC (Deepfake Detection Challenge) preview dataset 18 : This dataset contains 1,203 face videos of celebrities, including 408 real videos collected from YouTube with subjects of different ages, ethic groups and genders, and 795 deepfake videos synthesised from these real videos. Celeb-DF v1 35 [ ]: This dataset contains 6,229 face videos of celebrities, including 590 real videos collected from YouTube with subjects of different ages, ethic groups and genders, and 5,639 deepfake videos synthesised from these real videos. Celeb-DF v2 40 አግኙን ]: This dataset contains 3,363 face videos, covering 28 subjects, gender, and skin colour. It was created as a joint effort between two units of Google, Inc.: Google AI እና JigSaw . DeepFake Detection (DFD) Dataset 20 36 37 አግኙን ]: This dataset contains 60,000 indoor face videos (with 17.6 million frames) generated by face swapping, covering 100 subjects, four skin tones (white, black, yellow, brown), two gen-ders (male and female), different age groups (20-45), 26 nationalities, 7 different angles, 8 face expressions, and different head poses. DeeperForensics-1.0 27 [ ይህ ውሂብ ስብስብ በ 960 ጥያቄዎች በ 128,154 የግል ቪዲዮዎችን ያካትታል, ከ 3,426 ተስማሚ ተጫዋቾች ከ 23,654 የግል ቪዲዮዎች እና ከ 104,500 የግል ቪዲዮዎች በ 8 የተለያዩ ዘዴዎች (DF-128, DF-256, MM/NN face swap, NTH, FSGAN, StyleGAN, refinement, and audio swap) በመጠቀም የተፈጠረ ነው. DFDC (Deepfake Detection Challenge) full dataset 18 10 [ ]: This dataset contains 10,000 high-quality forgery videos, with video- and face-level annotations. The dataset focuses on a more challenging case for forgery detection: each video involves one to 15 individuals, but only some (a minority of) faces are manipulated. FFIW K (Face Forensics in the Wild) dataset 79 የ FFIW [ ]: This dataset contains 37,942 videos of paid subjects (395 Koreans and 8 Southeastern Asians), including 62,166 real videos and 175,776 fake ones created using six methods – FaceSwap, DeepFaceLab, FSGAN, First Order Motion Model (FOMM), Audio-driven Talking Face HeadPose (ATFHP) and Wav2Lip. The videos cover a balanced gender ratio and a wide range of age groups. Korean DeepFake Detection Dataset (KoDF) 36 [ ]: This dataset contains 1,737 videos with 1,666,816 frames, including 1,339,843 real frames and 326,973 fake frames generated using the Deep Video Portraits (DVP) [ ] method. The original videos were obtained from three sources: the dataset used in [ ], the Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS) [ ], and YouTube. Most videos have a resolution of 1280×720. VideoForensicsHQ 23 34 33 42 [ ]: This dataset contains 7,314 face sequences extracted from 707 deepfake videos that were collected completely from the Internet. It covers diverse scenes, multiple persons in each scene and rich facial expressions. Different from other deepfake video datasets, WildDeepfake contains only face sequences not the full videos. This makes the dataset more like between an image dataset and a video one. We decided to keep it in the video category since the selection process was still more video-focused. WildDeepfake 81 4.3 Deepfake Audio/Speech Datasets Voice conversion (VC) is a technology that can be used to modify an audio and speech sample so that it appears as if spoken by a different (target) person than the original (source) speaker. Obviously, it can be used to generate deepfake audio/speech samples. The biennial Voice Conversion Challenge that started in 2016 is a major challenge series on VC. Datasets released from this challenge series are very different from other deepfake datasets: the deepfake data is not included in the original dataset created by the organisers of each challenge, but in the participant submissions (which are retargeted/fake utterances produced by VC systems built by participants). The challenge datasets also include the evaluation (listening-based) results of all submissions. Some fake utterances may be produced by DL-based VC systems, so we consider all datasets from this challenge series relevant for our purpose of this survey. 38 [ ]: The original dataset created by the challenge organisers was derived from the DAPS (Device and Produced Speech) Dataset [ ]. It contains 216 utterances (162 for training and 54 for testing) per speaker from 10 speakers. Participating teams (17) developed their own VC systems for all 25 source-target speaker pairs, and then submitted generated utterances for evaluation. At least six participating teams used DL-related techniques (LSTM, DNN) in their VC systems (see Table 2 of the result analysis paper (የእነዚህ ላይ የተመሠረተ ጥያቄዎች በእርግጠኝነት ሊሆን ይችላል). Voice Conversion Challenge 2016 database 62 47 39 [ ]: The original dataset created by the challenge organisers was also based on the DAPS dataset. It contains 116 utterances (81 for training and 35 for testing) per speaker from 12 speakers in two different tasks (called Hub and Spoke). Participating teams (23 in total, all for Hub and 11 for Spoke) developed their own VC systems for all 16 source-target speaker pairs, and then submitted generated utterances for evaluation. Comparing with the 2016 challenge, more participating teams used DL-related techniques (e.g., WaveNet, LSTM, DNN, CycleGAN, DRM – deep relational models, and ARBM – adaptive restricted Boltzmann machines) in their VC systems. Voice Conversion Challenge 2018 database 44 [ ]: This dataset is based on the Effective Multilingual Interaction in Mobile Environments (EMIME) dataset , a bilingual (Finnish/English, German/English, and Mandarin/English) database. It contains 145 utterances (120 for training and 25 for testing) per speaker from 14 speakers for two different tasks (with 4 × 4 and 4 × 6 source-target speaker pairs, respectively). Participating teams (33 in total, out of which 31 for Task 1 and 28 for Task 2) developed their own VC systems for all source-target speaker pairs, and then submitted generated utterances for evaluation. Comparing with the 2018 challenge, DL-based VC systems were overwhelmingly used by almost all participating teams (WaveNet and WaveGAN among the most used DL-based building blocks). Voice Conversion Challenge 2020 database 70 40 A major set of deepfake speech datasets were created for the (Automatic Speaker Verification Spoofing and Countermeasures) Challenge (2015-2021) ለ 2019 እና 2021 data sets contain speech data that can be considered deepfakes. ASVspoof 41 [ ]: This dataset is based on the Voice Cloning Toolkit (VCTK) corpus , a multi-speaker English speech database captured from 107 speakers (46 males and 61 females). Two attack scenarios were considered: logical access (LA) involving spoofed (synthetic or converted) speech, and physical access (PA) involving replay attacks of previously recorded bona fide recordings). For our purpose in this survey, the LA scenario is more relevant. The LA part of the dataset includes 12,483 bona fide (real) utterances and 108,978 spoofed utterances. Some of the spoofed speech data for the LA scenario were produced using a generative model involving DL-based techniques such as long short-term memory (LSTM) የኦሪጂናል ( ], WaveRNN [ ], WaveCycleGAN2 [ ]. Note that the challenge organisers did not use the term “deepfake” explicitly, despite the fact that the DL-generated spoofed speech data can be considered as deepfakes. ASVspoof 2019 Challenge database 67 42 43 50 28 58 [ ]: This dataset contains bona fide and spoofed speech data for the logical access (LA) task. The challenge is still ongoing and we did not find a detailed paper on the dataset, so cannot include more details other than its size (7.8 GB after compression). Although we did not see details of the generative algorithms used to produce spoofed speech data, we believe similar DL-based algorithms were used like for the 2019 challenge. ASVspoof 2021 Challenge – Logical Access Database 14 [ ]: In 2021, the challenge included an explicitly defined track on deepfake, but the task description suggests that the organisers of the challenge considered a broader definition of the term “deepfake” by looking at spoofing human listeners rather than ASV (Automatic Speaker Verification) systems. The size of the dataset is 34.5 GB after compression. ASVspoof 2021 Challenge – Speech Deepfake Database 15 በዩናይትድ ስቴትስ (የዩናይትድ ስቴትስ) እና በዩናይትድ ስቴትስ (የዩናይትድ ስቴትስ) ውስጥ በዩናይትድ ስቴትስ (የዩናይትድ ስቴትስ) እና በዩናይትድ ስቴትስ (የዩናይትድ ስቴትስ) በዩናይትድ ስቴትስ (የዩናይትድ ስቴትስ) በዩናይትድ ስቴትስ (የዩናይትድ ስቴትስ) በዩናይትድ ስቴትስ (የዩናይትድ ስቴትስ) በዩናይትድ ስቴትስ (የዩናይትድ ስቴትስ) በዩናይትድ ስቴትስ (የዩናይትድ ስቴትስ) በዩናይትድ ስቴትስ (የዩናይትድ ]. This dataset was created to demonstrate a proposed voice cloning method. It is relatively small, and contains 134 utterances, including 10 real ones, 120 cloned ones, and 4 manipulated ones. Another dataset was created by Google AI and Google News Initiative ይህ ውሂብ ስብስቦች በ 68 የቴክኒካዊ "ወይነቶች" በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ ነው. 5 44 4.4 Hybrid Deepfake ውሂብ ስብስቦች እነዚህ መረጃዎች በ DARPA Media Forensics (MediFor) ፕሮግራም የተፈጠረ ናቸው. for the 2020 OpenMFC . There are two GAN-generated deepfake datasets, one with more than 1,000 deepfake images and the other with over 100 deepfake videos. The datasets were made available to registered participants of the competition only. NIST OpenMFC (Open Media Forensics Challenge) Datasets 45 46 47 [ ]: This dataset is named as “a versatile benchmark for comprehensive forgery analysis”. It contains 2,896,062 images and 221,247 videos, including 1,457,861 fake images and 121,617 fake videos. The videos and images cover seven image-level and eight video-level manipulation approaches, 36 different types of perturbations and more mixed perturbations, and a large number of annotation labels (6.3 million classification labels, 2.9 million manipulated area annotations and 221,247 temporal forgery segment labels). The dataset is being used for supporting the Face Forgery Analysis Challenge 2021 at the SenseHuman 2021 (3rd Workshop on Sensing, Understanding and Synthesizing Humans) , co-located at the ICCV 2021 conference . ForgeryNet 25 48 49 50 4.5 A Deepfake Dataset Generator [ ]: This is not actually a dataset per se, but a system for producing large datasets more automatically, including generating deepfake datasets. One may argue the automatically generated datasets are fake since they are not produced from real-world scenes. DatasetGAN 74 4.6 Subjective Quality of Deepfakes in Different Databases As mentioned in Section , subjective quality evaluation is necessary to evaluate the realness, realisticness, and naturalness of deepfake media. While there has been very limited work on this topic, in 2020, Jiang et al. [ ] conducted a user study on realness of deepfake videos. They recruited 100 professional participants (most of whom are computer vision researchers), who were asked to evaluate the realness of 30 randomly selected videos from 7 deepfake video datasets (DeeperForensics-1.0, UADFV, DeepFake-TIMIT, Celeb-DF, FaceForensics++, Deep Fake Detection, and DFDC). Participants were asked to respond to the statement “The video clip looks real.” and gave scores following a five-point Likert scale (1 – clearly disagree, 2 – weakly disagree, 3 – borderline, 4 – weakly agree, 5 – clearly agree). 4.7 27 Table shows the results. Interestingly, we can see a huge difference between the realness levels of different datasets. What is probably quite surprising is that FaceForensics++, one of the most widely used deepfake datasets, has a very low MOS score and less than 9% of participants considered the 30 selected videos as real. 3 Table 3: Human-judged subjective quality (realness) of deepfake videos in 7 datasets. The MOS scores were not reported by , but calculated by us based on the raw data shown in Table 3 of [ ]. ጂያንግ እና ኤል. 27 4.7 Discussion: Datasets Among all deepfake image and video datasets, a significant majority are about face images and videos. This is not surprising since face swapping, face attribution manipulation, and fully synthesised face images are among the hottest topics within deepfake research and real-world applications. We hope more non-face deepfake image and video datasets can be produced to support a broader range of research activities on deepfake. በ TABLE ውስጥ የተወሰነ ጥራት ውጤቶች indicate that it is important to check realness of deep-fake media to support any performance evaluation or comparison. To ensure that the quality evaluation of datasets is fair, transparent and reliable, standard procedures need defining and a common pool of qualified human experts should be used. 3 Many authors of deepfake-related datasets attempted to classify such datasets into different generations. Chronologically speaking, we could broadly split such datasets into two generations: before 2019 and since 2019. Typically, datasets created before 2019 are relatively less advanced and smaller, while those created after 2019 tend to be larger, more diverse (i.e., covering more attributes), and of higher quality (i.e., produced by more advanced generative models). This can also be seen from the data in Table , in which the top two datasets (DeeperForensics-1 and Celeb-DF) fall within the new generation (2020), while others belong to the old generation. In addition to the two generations, a newer generation has also emerged in 2021: a number of very recent datasets started focusing on more realistic deepfakes (i.e., in the wild) or more specified areas of deepfakes (e.g., 10 focusing on multiple faces in the same video, and KoDF focusing on Korean faces). This trend shows that the deepfake research community has grown significantly in the past few years so that narrower topics have also started gaining attention and interest from some researchers. 3 FFIW K 5 Deepfake-Related Challenges, Competitions & Benchmarks This section reviews initiatives aiming to advance the state-of-the-art of detection and generation of synthetic or manipulated media (such as video, image and audio) via competitions or challenges open to the public, and ongoing benchmarks tackling specific problems. 5.1 Detection of Manipulated Media The Deepfake Detection Challenge (DFDC) was an initiative promoted by an AI and Media Steering Committee , including BBC, Facebook, Amazon, Microsoft and New York Times, and some universities around the world including the University of Oxford. The competition remained open from 5 September 2019 till 31 March 2020, and involved 3 stages. At first, the DFDC preview dataset was released. At a later stage, the DFDC full dataset was also made available to the 2,114 participants of the competition incorporating face and audio swap techniques for generation of deepfake content. At the final stage, the submitted models were evaluated using a test dataset (referred to as the “black box dataset”) of 10,000 videos which included deepfake videos. The best performance on the black box dataset had an accuracy of 65.18%, according to the released results [ ]. Submissions were ranked በዩናይትድ ስቴትስ ውስጥ በዩናይትድ ስቴትስ ( ). All top five ranked models (the winner had the lowest overall log loss) are available on GitHub. Results indicate how challenging the detection of deepfake is since the best accuracy was low and “ ”, according to Dolhansky et al. [ ]. Figure shows a screenshot of the leaderboard with the five finalists. The first top ranked model used MTCNN (Multi-tasked Cascaded Convolutional Network), the second used WS-DAN (Weakly Supervised Data Augumentation Network), and the third used the EfficientNetB7 architecture. Meta compiling the common themes observed in the winning models, they were: clever augmentations, architectures, and absence of forensics methods. Moving forward, they called for “ አግኙን 51 52 in-the-wild 22 53 13 many submissions were simply random 19 2 የፎቶዎች እና ቪዲዮዎችን ልክ ያካትታል መፍትሔዎች. የኮንክቶፕ, ምዕራፍ, እና ሌሎች ምልክቶች ያካትታል የ deepfake ማረጋገጫ ሞዴሎች ለማሻሻል መንገድ ሊሆን ይችላል. The Automatic Speaker Verification Spoofing And Countermeasures Challenge Workshop (ASVspoof) has been running biennially since 2015. This competition is organised by an international consortium that includes Inria and EURECOM (France), University of Eastern Finland, National Institute of Informatics (Japan), and Institute for Infocomm Research (Singapore). This year the ASVspoof challenge includes, for the first time, a sub-challenge focused on where the envisioned use case is an adversary trying to fool a human listener. The metric used for evaluating performance of submitted solutions (i.e., classifiers) is EER. Four baseline solutions (also called “countermeasures”), each using a different technique, were made available to participants with their corresponding EER metric values. The ASVspoof 2021 Speech Deepfake Database containing audio recordings with original and spoofed utterances has also been made available. The competition involves three phases : አንድ ልማት ደረጃ, አንድ ግምገማ ደረጃ, እና በኋላ ግምገማ ደረጃ; ቡድንዎች አንድ ልማት ደረጃ ወደ ቀጣይ ደረጃ እንዴት በመስመር ላይ እንዴት አይደለም. ተጨማሪ መረጃ በ 2021 ግምገማ ላይ የሚገኝ ነው የተለጠፈው ግምገማ ፕሮግራም [ ]. የኮምፒውተር መተግበሪያዎች የ EER እንደ የፈጠራ ውጤታማነት ግምገማ መተግበሪያ ማረጋገጫ ማረጋገጫ ማረጋገጫ መተግበሪያዎችን በይነገጽ ምክንያት በይነገጽ መተግበሪያ መተግበሪያዎች ላይ የተመሠረተ መሆኑን ያውቃሉ. ” by the ISO/IEC 19795-1:2021 standard. Despite the fact that only the 2021 ASVspoof competition contained a track explicitly related to deepfake, some data in the ASVspoof 2019 dataset (Logical Access task) used for the 2019 competition was generated using DL-based algorithms as mentioned in Section . We expect that this also holds for the ASVspoof 2021 dataset (Logical Access task). The ASVspoof 2019 competition used the EER as secondary metric; the primary performance metric used was the tandem detection cost function (t-DCF) [ ]. According to its evaluation plan [ ], t-DCF assesses the performance of the whole tandem system whereby “ ”. It is calculated according to Eq. ( ), where cm ( ) and cm( ) are, respectively, “ አግኙን 54 Speech DeepFake 55 56 13 EER reporting is deprecated 57 4 63 69 a CM [countermeasure] serves as a ‘gate’ to determine whether a given speech input originates from a bona fide (genuine) user, before passing it the main biometric verifier (the ASV system) 17 P s P s the miss rate and the false alarm rate of the CM system at threshold s For further information about Eq. ( ), including constants 1 and 2, please refer to the ASVspoof 2019 evaluation plan [ ]. 17 C C 69 An implementation of the t-DCF metric has been made available by the ASVspoof 2019’s organisers in Python and Matlab formats. 58 59 The Face Anti-spoofing (Presentation Attack Detection) Challenge የ IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2020) የ IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2020) የ IEEE/CVF International Conference on Computer Vision (ICCV 2021) የ 2021 IEEE/CVF International Conference on Computer Vision (ICCV 2021) የ IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2020) የ 2019 እና የ 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2020) የ 2019 እና የ 2020 IEEE/CVF Conference (CVPR 2020) የ 2021 IEEE/CVF International Conference on Computer Vision (ICCVV 2021) ጋር ተስማሚ ነው. የመጀመሪያው ደረጃ “የመሠረተ ልምድ” ነው; በ 2021 ሚያዝያ ውስጥ በ CASIA-SURF HiFiMask dataset በመጀመር was released to participants. The second phase is the “final ranking phase” (June 2021), when the competition ended. The competition adopted the following performance metrics for evaluation of the solutions submitted: attack presentation classification error rate (APCER), normal/bona fide presentation classification error rate (NPCER/BPCER), and average classification error rate (ACER), in accordance with the ISO/IEC 30107-3:2017 standard. Figure provides the leaderboard for the top three solutions. 60 61 62 63 64 3 The FaceForensics Benchmark is an ongoing automated benchmark for detection of face manipulation. The organisers of the benchmark made the FaceForensics++ dataset available for training. Manipulated videos (4,000 in total) were created using four techniques, i.e., two computer graphics-based approaches (Face2Face and FaceSwap) and two learning-based approaches (DeepFakes and Neural Textures). The deepfakes videos were generated using a slightly modified version of FaceSwap , and the Neural Textures videos were created using the approach proposed by Thies et al. [ ]. The benchmark test dataset is created from the collection of 1,000 images randomly selected from either the manipulation methods or the original videos [ ]. Participants have to submit results to the benchmark, rather then code like other competitions; this is illustrated in Figure . The outcome of a submission is illustrated in Figure , where the scores are a measure of accuracy (Eq. ( )). 65 66 61 56 4a 4b 8 The Open Media Forensics Challenge (OpenMFC, formerly DARPA MFC) is an annual image and video forensics evaluation aiming to facilitate development of multimedia manipulation detection systems. It has been organised annually starting from 2017 under the name of DARPA MFC. In 2020, the National Institute of Standards and Technology (NIST) initiated the as a new evaluation platform, based on their previous experiences with the DARPA MFC series, to make the participation more convenient for all researchers. In OpenMFC 2020, two deepfake-related tasks were included for the first time: Image GAN Manipulation Detection (IGMD) and Video GAN Manipulation Detection (VGMD). The organisers provided an image evaluation dataset for the IGMD task, containing 1,000 images from over 200 image journals , and a video evaluation dataset for the VGMD task, including over 100 test videos. Furthermore, they provided the datasets used in the previous MFC challenges as development datasets. The challenge is composed of two main phases for development and evaluation, respectively, and a pre-challenge phase for quality control testing. For evaluation of submissions, AUC-ROC is used as the primary metric. Furthermore, CDR@FAR, where CDR refers to correct detection rate or TPR (Eq. ( )) and FAR refers to false alarm rate or FPR (Eq. ( )), is also used as a metric [ ]. The DeeperForensics Challenge 2020 is a deepfake face detection challenge held at the 2020 ECCV 67 68 OpenMFC 69 70 4 5 49 71 SenseHuman Workshop . The challenge used the DeeperForensics1.0 dataset. 72 የ "የመሠረተ ደረጃ" በ 2020 መጋቢት ውስጥ በመጀመር በ 100 ውጤታማ ክፍሎች ሊሆን ይችላል, እና በ 2020 መጋቢት በመጀመር በ 2020 መጋቢት በመጀመር በ 2 ውጤታማ ክፍሎች ሊሆን ይችላል. ), where is the number of videos in the hidden test set, is the ground truth label of video (fake:1, real:0), and ( ) ይህ ቪዲዮ የሚወዱት ምናልባት ነው is fake. 18 N yi i p yi i Results of the competition were discussed by Jiang et al. [ ]. The top solution used three models, i.e., EfficientNet-B0, EfficientNet-B1 and EfficientNet-B2, for classification. The second top used EfficientNet-B5 for both an image-based model and a video-based model. The third ranked solution used a 3D convolutional neural network (3DCNN). 73 26 The Face Forgery Analysis Challenge 2021 is a competition hosted at the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2021). It is organised by researchers from a number of organisations in China including universities and SenseTime Research (the research arm of SenseTime , one of the major AI “unicorns” in China). The challenge aims to advance the state-of-the-art in detection of photo-realistic manipulation of images and videos. Participants are able to use a large annotated face dataset (i.e., the ForgeryNet dataset) that was obtained by applying a number of techniques for manipulation (15) and perturbation (36) to train their solutions. The phases comprise of Forgery Image Analysis, Forgery Video Analysis, Forgery Video Temporal Localization phases, and the final phase (i.e., “private test”) where participants’ models will be tested against an unseen dataset. The following metrics will be used [ ]: AUC, average precision (AP) at some “temporal Intersection over Union” (AP@tIoU) compared to a threshold ∈ [0*. ,* 0*.*95], and average recall (AR) at (AR@ ) where is the top labels returned for multi-class classifiers. 74 75 25 t 5 K K K K The 2020 CelebA-Spoof Face Anti-Spoofing Challenge was hosted at the 16 የአውሮፓ የኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ቪዛ ኮምፒውተር ” [ ]. The organisers made the face CelebA-Spoof dataset available for the competition containing rich annotation across a range of attributes. The competition only had one phase where participants submitted their solutions to be evaluated against a test dataset; the spoof class was considered as “positive” and the live class as “negative”. Metric TPR@FPR was used and collected at three points where the TPR when FPR = 10 4 determined the final ranking. The top three finalists (see Figure ) የ Deep Learning ሞዴሎች ResNet, EfficientNet-B7, እና Central Difference Convolutional Networks (CDCN) እና Dual Attention Network (DAN) አጠቃቀም አንድ አዲስ መዋቅር አጠቃቀም. ሁለቱም ከፍተኛ-ጥራት መፍትሔዎች የእነዚህ ሞዴሎች አፈጻጸም ለማሻሻል የተለያዩ ፕሮግራሞች አጠቃቀም: ከፍተኛ-ጥራት መፍትሔ በ heuristic voting scheme ጥቅም ላይ ነበር, እና ሁለተኛ-ጥራት መፍትሔ በ weight-after-sorting Strategy ጥቅም ላይ ነበር. 76 የ whether a presented face is live or spoof 76 − 5 The 2021 CSIG Challenge is the second edition of a challenge organised by the China Society of Image and Graphics . The 2021 challenge has the Fake Media Forensic Challenge as its 6 track, co-organised by CSIG’s Digital Media Forensics and Security Technical Committee and Institute of Information Engineering, Chinese Academy of Sciences . This track has two tasks, one on deepfake video detection, and the other on deepfake audio/speech detection. For the deepfake video detection task, the dataset used contains a public training set with 10,000 sound-free face videos (including 4,000 fake videos), a public test set with 20,000 face videos (the percentage of deepfake videos is unknown to participants), and a private test set that will be determined and used at the final session for selecting the winners. All videos contain faces of Eastern Asian people, and cover a wide range of parameters such as multiple resolutions and encoding quality factors, the use of blurring or sharpening filters, and added noise. Deepfake videos were created using public tools including DeepFaceLab [ ], Faceswap , Faceswap-GAN, Recycle-GAN [ ] and ALAE (Adversarial Latent Autoencoders) [ ]. For the deepfake audio/speech detection task, the dataset used contains a public training set with 10,000 speech samples (including 6,000 fake ones), a public test set with 20,000 face videos (the percentage of deepfake videos is unknown to participants), and a private test set for the final session (the same as the deepfake video detection task). The tools used for generating the fake speech samples include TTS (text-to-speech) voice synthesis tools and VC (voice conversion) tools. The main TTS tools used include open-source tools such as DeepVoice, TensorFlowTTS and GAN-TTS [ ] and commercial software tools such as those from iFlytek and IBM. The main VC tools used include Adaptive-VC and CycleGAN-VC [ ]. For both deepfake detection tasks, the performance metric used is log loss. 77 78 79 የ 80 81 53 82 6 54 83 8 84 29 2020 China Artificial Intelligence was the second edition of a Chinese AI competition open for the general public to participate, organised by the municipal government of the City of Xiamen in China. In 2020, it had two sub-competitions, Multimedia Information Recognition Technology Competition and Language and Knowledge Technology Competition . The Multimedia Information Recognition Technology Competition included two tasks on deepfakes: one on deepfake video detection and one on deepfake audio/speech detection . The deepfake video detection task used 3,000 videos, and log loss was used as the sole performance metric. The deepfake audio/speech detection task used 20,000 audio samples (mostly in Chinese, and the remaining in English), and EER was used as the sole performance metric. For both tasks, the ratio between real and deepfake samples was 1:1. We did not find where to download the datasets used for the tasks nor a more detailed technical description of the datasets. For the deepfake video detection tasks, the top two winning teams (with an A prize) were from Netease (Hangzhou) Network Co., Ltd. and Beijing RealAI Technology Co., Ltd., followed by three other teams winning a B prize: Xiamen Fuyun Information Technology Co., Ltd.; Institute of Computing Technology, Chinese Academy of Sciences; and Wuhan Daqian Information Technology Co., Ltd. For the deepfake audio/speech task, there was no team winning an A prize, but one team winning a B prize: SpeakIn Technologies Co., Ltd. The final results of some teams were published, but some teams were allowed to hide their results. We did not find a detailed technical report summarising the results and explaining the work of the winning teams. 85 86 87 88 89 One of the B-prize winning team is from Beijing RealAI Technology Co., Ltd., a Chinese company active in deepfake-related R&D. 5.2 Generation of Manipulated Media The Voice Conversion Challenge is a biennial competition that has been running since 2016. The challenge and the corresponding workshop, hosted at the INTERSPEECH conference , is supported by the SynSig (Speech Synthesis Special Interest Group) of the International Speech Communication Association (ISCA) . Its aim is to promote progress in voice conversion (VC) technology that can be applied to a number of positive and negative use cases, such as spoofing voice biometric systems. The 2020 challenge focused on speaker conversion, a sub-problem of VC, and included two tasks. For the first task “intra-lingual semi-parallel voice conversion”, participants had to develop 16 VC systems (speaker-pair combinations) including male and female speakers and English sentences, using the provided Voice Conversion Challenge 2020 database v1.0 for training (refer to Section ). For the second task “cross-lingual voice conversion”, participants had to develop 24 VC systems, also including male and female speakers, but uttering sentences in three languages (Finnish, German and Mandarin), based on the provided training dataset. Figure illustrates the process of training and generation of VC systems. 90 91 92 93 4 6 Submissions were evaluated for “ ” . As such, the organisers used አግኙን ] and recruited both native and non-native English speakers (i.e., Japanese native speakers) via crowd-sourcing for the listening tests. Naturalness (answering the question “How natural does the converted voice sound? ”) was measured using the metric MOS (covered in Section 4.6), and similarity (answering the question “how similar the converted voice sound comparing source and target speakers? ”) was measured in terms of speaker recognition as “same” or “different”, as elaborated by Wester et al. [68]. Tests also focused on the effects of language differences on the performance of VC systems submitted to the competition. The most popular CNN/RNN/GANbased VC systems submitted used WaveNet, WaveRNN, and Parallel WaveGAN. Results indicated that, in terms of similarity, the best performing VC systems were as good as natural speech but none reached human-level naturalness for task 1; scores were lower for task 2 which was more complex [70]. The organisers of the 2020 competition also used objective evaluation [12]. The metrics used for evaluation of speaker similarity were: equal error rate (EER), false acceptance rate of target (P tar fa ), miss rate of source (P src miss), and cosine similarity of speaker embedding vectors (cos-sim) according to Eq. (19) where A is the speaker embedding vectors for the converter audio and B is the speaker embedding vectors for the original audio. The performance of the VC systems as a spoof countermeasure was also evaluated using EER, while to evaluate the quality of the subjective MOS obtained via listening tests, a DL-based model to predict MOS, called MOSNet [43], was used. Lastly, to evaluate intelligibility of the converted transcribed speech, in comparison with the original transcribed speech, the word error rate (WER) [4] was used. WER is calculated according to Eq. (20) where I refers to insertions, D refers to deletions, S refers to substitutions, and N refers to the total number of words in the original transcript. perceived naturalness and similarity through listening tests 94 subjective evaluation 70 The Deepfake Africa Challenge (2021) is a new initiative of the AI Africa Expo, in partnership with a film and media production company (Wesgro) and the African Data Science competition platform Zindi. Its aim is “ ”. Eligible participants were required to be citizens and residents of the African continent. Submissions, accepted up to end of July 2021, can be either video or audio. Evaluation of submissions is defined in terms of artistic creativity, relevance of challenge topic, and innovation in the process of generation as long as participants use tools and packages publicly available. The top three finalists will receive a prize, present their work at the Expo, and will have to grant copyrights to Zindi. Unlike the other competitions reviewed in this section, which were focused on advancing the state-of-the-art in detection of synthetic or manipulated media, this competition focused on the generation of deepfake which seems more humanities-centred. This is a trend observed in arts [ ] and culture [ ].5.3 Generation and Detection of Manipulated Media 95 to create convincing deepfakes to highlight the power of this synthetic media, illustrating its creative potential for exploitation for both positive and negative outcomes and focusing debate about its ethical use / misuse in an African context 31 57 The DeepFake Game Competition (DFGC) is in its first edition, hosted at the 2021 International Joint Conference on Biometrics (IJCB 2021). Its organisers are mainly from the Institute of Automation Chinese Academy of Sciences (CASIA). The idea of the competition was to promote an adversarial game between agents pushing for advances in both deepfake creation and detection. In order to achieve this, a 6-stage protocol was designed interleaving three creation phase (C-phase) and detection phase (D-phase), typically one week apart; submissions closed in April 2021. Both C-phases and D-phases were bound to the Celeb-DF (v2) dataset [ ], containing 6,229 videos (590 real/original videos and 5,639 fake/manipulated videos), for training purposes. As such, submissions to a C-phase would consist of datasets extracted from Celeb-DF (v2) which included novel face-swap approaches to obtain evaluation results. Submissions to a D-phase would consist of detection models/codes to obtain evaluation results. The models submitted for a D-phase were evaluated against the datasets submitted for the previous C-phase [ የክፍያ ማረጋገጫ ላይ ጥቅም ላይ የሚጠቀሙትን ማረጋገጫዎች were: a detection score, used for evaluation of a D-phase, and a creation score, used for evaluation of a C-phase. The top three finalists for the detection phase employed CNN-based classifiers EfficientNet-B3, Efficientnet-B0 and EfficientNetV2. 96 40 52 97 The Detection Score ( ) metric captures the models’ ability to correctly classify fake images submitted to the previous C-phase against a set of real images in the CelebDF test dataset. It is calculated using Eq. ( ), where is the number of valid submissions of created synthesis test sets in the last C-phase. DS 21 NC The Creation Score ( ) metric used to evaluate creation models submitted to this challenge is calculated by Eq. ( ), where is the number of valid submissions of detection methods in the last D-phase, the noise score ( noise) penalises noisy images, the other three parts of the equation relate to the following : “ ”. CS 22 ND S 98 ID level similarity to the donor ID, image level similarity to the target frame, and the deception ability against detection models. ID level similarity is scored by a face recognition model using dot product of two ID features (fake face ID and donor ID). The image level similarity is scored by SSIM [Structural Similarity Index] to make sure the face-swapped image is similar to the corresponding target image in content and quality Peng et al. [ ] observed a commonality between the three winning teams for the creation task, i.e., the use of the FaceShifter [ ] framework for face swapping. They highlighted two overall reflections about the competition: (1) the limited diversity of the deepfake datasets submitted and the use of repetitive methods to generate them, and (2) the limited size of the Celeb-DF (v2) dataset itself flagging the need for a larger dataset for next year’s competition. The organisers of the competition also applied the top two detection models to unseen datasets (DFDC and FaceForensics++) and noticed that they do not generalise well. 52 37 6 A Meta-Review of Deepfake-Related Surveys This section presents a meta-review of 12 selected deepfake-related survey papers, including eight published in English [ , , , – , , ] and four published in Chinese [ , , , ]. It covers the following aspects in a systematic manner: definitions and scope, performance metrics, datasets, challenges/competitions/benchmarks, performance comparison, key challenges and recommendations. 16 45 46 64 66 71 73 7 38 41 59 The meta-review aims at drawing some high-level insights for monitoring future development of deepfake-related technologies and their applications. 6.1 Definitions and Scope As we discussed in Section , among researchers, practitioners and law makers there is no universally accepted definition of “deepfake” as a term. This is also reflected in how the authors of the 12 survey papers considered this aspect. Most authors talked about the history of deepfakes and pointed out that the term reflects the combination of “deep learning” and “fake”, but some used a broader definition, e.g., Lyu [ ] defined deepfake as “ ”. Some authors also referred to deepfake-related legislations, but none of them pointed out that the definitions in some such legislations are completely different from the more technical definitions involving the use of deep learning. No authors discussed the blurred boundary between deepfakes and non-deepfakes, although some surveys actually cover both, e.g., Tao et al. [ [ ] በጽሑፍ ትክክለኛነት ላይ የተመሠረተ እና በጽሑፍ ላይ በጽሑፍ ላይ የተመሠረተ "የብዙ ትክክለኛነት" አይደለም. 1.1 45 high quality fake videos and audios generated by AI algorithms 59 In terms of the scope, while some authors (correctly) considered all types of media that can be produced by deepfake-related techniques [ , , , ], some considered only a narrow scope, e.g., authors of [ , , , ] considered only videos, and only authors of [ , ] have considered images and videos. Another phenomenon we observed is that many authors focused more on face images and videos, and authors of three surveys [ አግኙን , ] even limited the definition of “deepfake” to such a narrow scope: 38 41 45 65 7 64 71 73 16 66 16 64 71 Deshmukh and Wankhade [ ] defined it as “ ”; 16 a technology which creates fake images or videos of targeted humans by swapping their faces [by] another character saying or doing things that are not absolutely done by them and humans start believing in such fake as it is not always recognisable with the everyday human eye Younus and Hasan [ ] considered deepfake as a technique allowing “ ”; and 71 any computer user to exchange the face of one person with another digitally in any video Tolosana et al. [ ] defined it as “ ”. 64 a deep learning based technique able to create fake videos by swapping the face of a person by the face of another person እንደዚህ የሚፈልጉት ጥንካሬ ትዕዛዞች እና ስኬትዎች በይነገጽ ሊሆን ይችላል እና በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ በይነገጽ ይረዳል. We call on more researchers to accept a broader definition of “deepfake” so that highly realistic/natural media of any kind generated by a sophisticated automated method (often AI-based) is considered deepfake. Here, we provide two examples of such a broader definition: the image2image (or pixel2pixel) technique [ ] that allows the production of deepfake images and videos of any objects (e.g., the “horse2zebra” deepfake image shown in Figure ), and the the so-called “deepfake geography [ [ ]”, በ AI ላይ የተመሠረተ ቴክኖሎጂዎች ለመፍጠር እውነተኛ ተንቀሳቃሽ ፎቶዎችን ለመፍጠር ይጠቀማሉ. 80 7 77 Another important fact missed or not sufficiently discussed by authors of all the 12 surveys is that deepfake techniques can be used for positive applications, e.g., creative arts, entertainment and protecting online users’ privacy. We call for more researchers and practitioners to follow the proposal in the 2020 Tencent AI White Paper [ ] to start using the more neutral-sounding term “deep synthesis”. Accordingly,we can use different words for different types of data generated using “deep synthesis” techniques, e.g., “deep art”, “deep animation”, “deep music”, and “deepfake”. While authors of the 12 survey papers did not recognise the positive applications of “deepfake” technologies, some other researchers did, e.g., organisers of the Voice Conversion Challenge 2020 who said the VC technology (for speech deepfake) “ ”. 60 99 is useful in many applications, such as customizing audio book and avatar voices, dubbing, movie industry, teleconferencing, singing voice modification, voice restoration after surgery, and cloning of voices of historical persons 6.2 Performance Metrics በተሳካ ሁኔታ, የ 12 ግምገማዎች ማንኛውምም በግልጽ አፈጻጸም ሜትሪኮች ያካትታል. አንዳንድ በግልጽ አፈጻጸም ሜትሪኮች ያካትታል እና በግልጽ የ Deepfake Generation እና Detection Methods አፈጻጸም ያካትታል. በአጠቃላይ የተጠቃሚ አፈጻጸም ሜትሪኮች ትክክለኛነት, ERR, እና AUC ያካትታል. ይህ እንደዚህ ግምገማዎች ላይ ገጾች ጥንካሬዎች ሊተካል ይችላል, ይህም የፈጠራ ሜትሪኮች በተሳካ ሁኔታ ያካትታል. የ Deepfake ቁልፍ ጥራት በከፍተኛ ፍጹም ጥንካሬ Generation እና በከፍተኛ ፍጹም ጥንካሬ ጥንካሬ ጥንካሬ ጥንካሬ ጋር የተገናኙ ሊሆን ይችላል 6.3 Datasets Many of the 12 survey papers list a number of deepfake-related datasets, but none of them have coverage as complete as ours shown in Section . For instance, none of the surveys have covered the Voice Conversion Challenge 2016/2018/2020 datasets and the ASVspoof 2019/2021 datasets are covered briefly only in two surveys [ , ]. In addition, more recent deepfake datasets especially those released in 2021 are also not covered by any of the surveys. We believe that our Section is the most comprehensive review of deepfake-related datasets so far. 4 38 59 4 አንዳንድ ግምገማ ወረቀቶች በጣም ታዋቂ ውሂብ ስብስቦች ያካትታል, ለምሳሌ, Verdoliva [ ፎቶዎች በይፋ መግዛት ወይም የ AI-based methods በይፋ መግዛት አይሆንም, እና አንዳንድ ግምገማዎች (እን. ]) mentioned ASVspoof 2015 datasets but we did not see the use of deep learning for generating data used in the dataset. 66 38 6.4 ግምገማዎች, ክወናዎች እና ምናሌዎች Many surveys cover deepfake-related challenges, competitions and benchmarks. The coverage is, however, mostly limited, and some challenges (e.g., the Voice Conversion Challenge 2016/2018/2020 and the two Chinese challenges we covered in Section የፈተናዎች, የኮምፒውተር እና የኮምፒውተር ግምገማዎች ዝርዝር ዝርዝር ዝርዝር ዝርዝር ዝርዝር ዝርዝር ዝርዝር ነው. . Similar to the datasets we covered in Section , we believe that our coverage of deepfake-related challenges, competitions and benchmarks in Section is also the most comprehensive so far. 5 5 4 5 6.5 አፈጻጸም ተመሳሳይነት Most surveys have a good coverage of related methods for deepfake generation and detection, but only some explicitly covered performance comparison between different methods [ , , ]. 38 46 64 ከሁሉም ግምገማዎች መካከል, Li et al. [ ] conducted the most comprehensive study on performance of different deepfake detection methods. In addition to showing the performance metrics of a number of deepfake detection methods in Table 3 of [ [ ], እነርሱ ደግሞ በቴሌት ውስጥ የተያያያዙ እንደ የተለያዩ ዓይነት የ deepfake ማረጋገጫ መሳሪያዎች የጋራ ባህሪያት እና ጥያቄዎች ይመልከቱ. . Furthermore, they also looked at research on robustness of deepfake detection methods against adversarial samples, referring to some work that showed a lack of such robustness. 38 38 4 Due to quality issues of many deepfake-related datasets (discussed in Section ), we need to treat any performance metrics and comparison of different detection methods with caution. Without testing all methods on a sufficiently large, diverse and high-quality deepfake dataset, the performance comparison results can be misleading. This highlights the importance of having more challenges, competitions and benchmarks to encourage performance comparison on standard datasets and using consistent performance metrics. 4.6 6.6 Challenges and Recommendations አንዳንድ ግምገማዎች መተግበሪያዎች ለ deepfake ማህበረሰብ አንዳንድ አስፈላጊ ልምድ እና የመጨረሻ ምርምር መመሪያዎችን ይሰጣሉ. Not surprisingly, how to develop more robust, scalable, generalisable and explainable deepfake detection methods is one of the most discussed key challenges and also a major future research direction [ , , , , , , , , ]. Considering the arms race between deepfake generation and detection, this research direction will likely remain the hottest topic in deepfake research. 7 16 38 41 45 59 65 66 71 A couple of surveys [ , ] mentioned fusion as a key future research direction, where “fusion” refers to combining different methods (e.g., combining multiple detectors of different types) and data sources (e.g., jointly considering audio-visual analysis) to achieve better performance for deepfake detection. Lyu [ ] suggested that, for detection of deepfake videos, we need to consider video-level detection more, which can be considered fusion of detection results of all video frames. 38 66 45 The authors of three surveys, Lyu [ ] , Deshmukh and Wankhade [ ] and Younus and Hasan [ ], argued that better (higher-quality, more up-to-date, and more standard) deepfake datasets are needed to develop more effective deepfake detection methods. Lyu [ ] also suggested that we need to consider effects in training data and improve the evaluation of datasets. We agree with them on these points. 45 16 71 45 social media laundering Tao et al. [ ] suggested that low-cost deepfake generation/detection should be considered as a future research direction. This is a valid recommendation since lightweight methods will allow less powerful computing devices (e.g., IoT devices) to benefit from such technologies. 59 Two Chinese surveys [ አግኙን ] also mentioned the need to have new deepfake-related legislations combating malicious use of deepfakes and the need to train end users such as journalists. This is likely an area where interdisciplinary research can grow. 38 41 There are also other ad-hoc recommendations given by the authors of some surveys. For example, Lyu [ ] argued that deepfake detection should be considered a (more complicated) multi-class, multi-label and local detection problem. Tolosana et al. [ [ ] የተለያዩ ጥቁር-fake Generation methods (እንጨት መተግበሪያ, Identity swap, Attribute Manipulation, እና Expression swap) ለ ልዩ ምርምር መስፈርቶች ይመልከቱ. [ Liang et al. ] ] and Li et al. [ የዲጂታል watermarking እና blockchain ቴክኖሎጂዎችን በመጠቀም እንደ የዲጂታል ውሃ ማረጋገጫ እና የዲጂታል ውሃ ማረጋገጫ እና የዲጂታል ውሃ ማረጋገጫ እና የዲጂታል ውሃ ማረጋገጫ እና የዲጂታል ውሃ ማረጋገጫ እና የዲጂታል ውሃ ማረጋገጫ እና የዲጂታል ውሃ ማረጋገጫ እና የዲጂታል ውሃ ማረጋገጫ እና የዲጂታል ውሃ ማረጋገጫ እና የዲጂታል ውሃ ማረጋገጫ እና የዲጂታል ውሃ ማረጋገጫ እና የዲጂታል የዲጂታል ውሃ ማረጋገጫ እና የዲጂታል ውሃ ማረጋገጫ እና የዲጂታል የዲጂታል ውሃ ማረጋገጫ እና የዲጂታል ውሃ ማረጋገጫ ማረጋገጫ ማረጋገጫ ማረጋገጫ ማረጋገጫ ማረጋገጫ ማረጋገጫ ይሰጣል. 45 64 41 38 7 Conclusion The rapid growth in the capability to manipulate media or create synthetic media which look realistic and natural paved the way for deepfakes. At first, this paper adopted a critical approach to look at different definitions of the term “deepfake”. In that regard, we point out the different contradicting definitions and call for the wider community to consider how to define a new term that has a more consistent scope and meaning. For instance, replacing “deepfake” by “deep synthesis” can be more inclusive by embracing positive applications of deepfake techniques, e.g., in entertainment and for simulation purposes. This paper provided a comprehensive overview of multiple aspects of the deepfake ecosystem drawing from the research literature and other online sources published in two languages: English and Chinese. It covers commonly used performance metrics and standards, related datasets, challenges, competitions and benchmarks. It also presents a meta-review of 12 selected deepfake-related survey papers published in 2020 and 2021, covering not only the above mentioned aspects, but also highlighting key challenges and recommendations. አስተያየቶች [1] Darius Afchar, Vincent Nozick, Junichi Yamagishi, and Isao Echizen. 2018. MesoNet: A Compact Facial Video Forgery Detection Network. In . IEEE, 1–7. Proceedings of the 2018 IEEE International Workshop on Information Forensics and Security https://doi.org/10.1109/WIFS.2018.8630 761 [2] Henry Ajder, Giorgio Patrini, Francesco Cavalli, and Laurence Cullen. 2019. The State of Deepfakes: Landscape, Threats, and Impact. Deeptrace. , 27 pages. https://sensity.ai/reports/ [3] Zahid Akhtar እና Tiago H. Falk. 2017 የኦዲዮ-ቪዲዮ ሜልሚዲያ ጥራት ግምገማ: ሙሉ በሙሉ ግምገማ. 5 (2017), 21090–21117. IEEE Access https://doi.org/10.1109/ACCESS.2017.2750918 [4] Ahmed Ali and Steve Renals. 2018. Word Error Rate Estimation for Speech Recognition: e-WER. In . Association for Computational Linguistics, 20–24. Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics https://doi.org/10.18653/v1/P18-2004 [5] Sercan O ̈ . Arık, Jitong Chen, Kainan Peng, Wei Ping, እና Yanqi Zhou. 2018. የኒውሮል ስልክ ክሎኒንግ ከሁለቱም ሞዴሎች ጋር. በ . Curran Associates Inc., 10040–10050. Proceedings of the 32nd International Conference on Neural Information Processing Systems https://papers.nips.cc/paper/201 8/hash/4559912e7a94a9c32b09d894f2bc3c82-መጽዕኖ.html [6] Aayush Bansal, Shugao Ma, Deva Ramanan, እና Yaser Sheik. 2018. Recycle-GAN: የማይታመን ቪዲዮ መተግበሪያ. በ . Springer, 17 pages. Proceedings of the 2018 European Conference on Computer Vision https://doi.org/10.1007/978-3-030-01228-1 8 [7] Yu-xuan Bao, Tian-liang Lu, and Yan-hui Du. 2020. Overview of Deepfake Video Detection Technology. 47, 9 (2020), 283–292. Computer Science https://doi.org/10.11896/jsjkx.200400130 [8] Mikol-aj Bin´kowski, Jeff Donahue, Sander Dieleman, Aidan Clark, Erich Elsen, Norman Casagrande, Luis C. Cobo, and Karen Simonyan. 2019. High Fidelity Speech Synthesis with Adversarial Networks. https://doi.org/10.48550/ARXIV.1909.11646 [9] ![](file:///C:/Users/user/AppData/Local/Temp/msohtmlclip1/01/clip_image050.gif)Madeline Brady. 2020. Deepfakes: A New Desinformation Threat? Report by the Democracy Reporting International. , 9 pages. https://democracy-reporting.org/dri publications/de epfakes-a-new-disinformation-threat/ [10] Umur Aybars Ciftci, Ilke Demir, and Lijun Yin. 2020. FakeCatcher: Detection of Synthetic Portrait Videos using Biological Signals. (2020), 17 pages. IEEE Transactions on Pattern Analysis and Machine Intelligence https://doi.org/10.1109/TPAMI.2020.3009287 [11] Hao Dang, Feng Liu, Joel Stehouwer, Xiaoming Liu, and Anil K. Jain. 2020. On the Detection of Digital Face Manipulation. In . IEEE, 10 pages. Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition https://doi.org/10.1109/CVPR42600.2020.00582 [12] Rohan Kumar Das, Tomi Kinnunen, Wen-Chin Huang, Zhen-Hua Ling, Junichi Yamagishi, Zhao Yi, Xiaohai Tian, and Tomoki Toda. 2020. Predictions of Subjective Ratings and Spoofing Assessments of Voice Conversion Challenge 2020 Submissions. In . International Speech Communication Association, 99–120. የ Blizzard Challenge እና የ Voice Conversion Challenge 2020 ተመራማሪዎች https://doi.org/10.21437/VCC BC.2020-15 [13] H´ector Delgado, Nicholas Evans, Tomi Kinnunen, Kong Aik Lee, Xuechen Liu, Andreas Nautsch, Jose Patino, Md Sahidullah, Massimiliano Todisco, Xin Wang, and Junichi Yamagishi. 2021. ASVspoof 2021: Automatic Speaker Verification Spoofing and Countermeasures Challenge Evaluation Plan. https://www.asvspoof.org/asvspoof2021/asvspoof2021 evaluation plan.pdf [14] H´ector Delgado, Nicholas Evans, Tomi Kinnunen, Kong Aik Lee, Xuechen Liu, Andreas Nautsch, Jose Patino, Md Sahidullah, Massimiliano Todisco, Xin Wang, and Junichi Yamagishi. 2021. ASVspoof 2021 Challenge - Logical Access Database. https://doi.org/10.5281/zenodo.4837263 [15] H´ector Delgado, Nicholas Evans, Tomi Kinnunen, Kong Aik Lee, Xuechen Liu, Andreas Nautsch, Jose Patino, Md Sahidullah, Massimiliano Todisco, Xin Wang, and Junichi Yamagishi. 2021. ASVspoof 2021 Challenge - Speech Deepfake Database. https://doi.org/10.5281/zenodo.4835108 [16] ![](file:///C:/Users/user/AppData/Local/Temp/msohtmlclip1/01/clip_image051.gif)Anushree Deshmukh and Sunil B. Wankhade. 2021. Deepfake Detection Approaches Using Deep Learning: A Systematic Review. In . Springer, 293–302. Intelligent Computing and Networking: Proceedings of IC-ICN 2020 (Lecture Notes in Networks and Systems, Vol. 146) https://doi.org/ 10.1007/978-981-15-7421-4 27 [17] Xinyi Ding, Zohreh Raziei, Eric C. Larson, Eli V. Olinick, Paul Krueger, እና Michael Hahsler. 2020 በከፍተኛ ማወቅ እና በባህላዊ ግምገማዎችን በመጠቀም የጽሑፍ ማወቅ. 2020, 1 (2020), 1–12. EURASIP Journal on Information Security https://doi.org/10.1186/s13635-020-00109-8 [18] Brian Dolhansky, Joanna Bitton, Ben Pflaum, Jikuo Lu, Russ Howes, Menglin Wang, እና Cristian Canton Ferrer. 2020. The DeepFake Detection Challenge (DFDC) Dataset. https://doi.org/10.48550/ARXIV.2006.07397 [19] Brian Dolhansky, Joanna Bitton, Ben Pflaum, Jikuo Lu, Russ Howes, Menglin Wang, and Cristian Canton Ferrer. 2020. The DeepFake Detection Challenge (DFDC) Dataset. arXiv:2006.07397. https://arxiv.org/abs/2006.07397 [20] Nick Dufour and Andrew Gully. 2019. Contributing Data to Deepfake Detection Research. Google AI Blog. https://ai.googleblog.com/2019/09/contributing-data-to-deepfake-detectio n.html [21] Ricard Durall, Margret Keuper, Franz-Josef Pfreundt, and Janis Keuper. 2019. Unmasking Deep-Fakes with Simple Features. https://doi.org/10.48550/ARXIV.1911.00686 [22] Cristian Canton Ferrer, Brian Dolhansky, Ben Pflaum, Joanna Bitton, Jacqueline Pan, and Jikuo Lu. 2020. Deepfake Detection Challenge Results: An Open Initiative to Advance AI. Meta AI Blog. https://ai.facebook.com/blog/deepfake-detection-challenge-results-an-open-initia tive-to-advance-ai/ [23] Gereon Fox, Wentao Liu, Hyeongwoo Kim, Hans-Peter Seidel, Mohamed Elgharib, እና Christian Theobalt. 2021. Videoforensicshq: ከፍተኛ ጥራት የተመሠረተ ቪዲዮዎችን ማረጋገጫ. በ . IEEE, 1–6. Proceedings of the 2021 IEEE International Conference on Multimedia and Expo https://doi.or አጠቃቀም / 10 1109 / ICME51207.2021.9428101 [24] Haiying Guan, Andrew Delgado, Yooyoung Lee, Amy N. Yates, Daniel Zhou, Timothee Kheyrkhah, and Jon Fiscus. 2021. User Guide for NIST Media Forensic Challenge (MFC) Datasets. https://doi.org/10.6028/NIST.IR.8377 [25] Yinan He, Bei Gan, Siyu Chen, Yichun Zhou, Guojun Yin, Luchuan Song, Lu Sheng, Jing Shao, and Ziwei Liu. 2021. ForgeryNet: A Versatile Benchmark for Comprehensive Forgery Analysis. In . IEEE, 4360–4369. Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition https://doi.org/10.1109/CVPR46437.2021.00434 [26] Liming Jiang, Zhengkui Guo, Wayne Wu, Zhaoyang Liu, Ziwei Liu, Chen Change Loy, Shuo Yang, Yuanjun Xiong, Wei Xia, Baoying Chen, Peiyu Zhuang, Sili Li, Shen Chen, Taiping Yao, Shouhong Ding, Jilin Li, Feiyue Huang, Liujuan Cao, Rongrong Ji, Changlei Lu, እና Ganchao Tan. 2021. DeeperForensics Challenge 2020 on Real-World Face Forgery Detection: Methods and Results. arXiv:2102.09471. https://arxiv.org/pdf/2102.09471.pdf [27] Liming Jiang, Ren Li, Wayne Wu, Chen Qian, and Chen Change Loy. 2020. DeeperForensics-1.0: A Large-Scale Dataset for Real-World Face Forgery Detection. In . IEEE, 2886–2895. Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition https://doi.org/ 10.1109/CVPR42600.2020.00296 አጠቃቀም [28] Nal Kalchbrenner, Erich Elsen, Karen Simonyan, Seb Noury, Norman Casagrande, Edward Lockhart, Florian Stimberg, Aaron van den Oord, Sander Dieleman, and Koray Kavukcuoglu. 2018. Efficient Neural Audio Synthesis. https://doi.org/10.48550/ARXIV.1802.08435 [29] Takuhiro Kaneko and Hirokazu Kameoka. 2017. Parallel-Data-Free Voice Conversion Using Cycle-Consistent Adversarial Networks. https://doi.org/10.48550/ARXIV.1711.11293 [30] Tero Karras, Samuli Laine, and Timo Aila. 2019. A Style-based Generator Architecture for Generative Adversarial Networks. In . IEEE, 4401–4410. የ 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (የኮምፒውተር ቪዛ እና ሞዴሎች ማረጋገጫ) https://doi.org/10.1109/CVPR.2019.00453 [31] Ross Kelly. 2021. Digital Art Exhibition to Showcase Creative Potential of AI. DIGIT News. https://www.digit.fyi/digital-art-exhibition-to-showcase-creative-potential-of-ai/ [32] Ali Khodabakhsh, Raghavendra Ramachandra, Kiran Raja, Pankaj Wasnik, and Christoph Busch. 2018. Fake Face Detection Methods: Can They Be Generalized?. In አዲሱ, 1 – 6 የ 2018 International Conference of the Biometrics Special Interest Group (የቢሮሜቲክስ ልዩ ፍላጎቶች ቡድን) https://doi.org/10.2 3919/BIOSIG.2018.8553251 በዩናይትድ ስቴትስ [33] Hyeongwoo Kim, Mohamed Elgharib, Hans-Peter Zoll¨ofer, Michael Seidel, Thabo Beeler, Christian Richardt, and Christian Theobalt. 2019. Neural Style-Preserving Visual Dubbing. 38, 6, Article 178 (2019), 13 pages. ACM Transactions on Graphics https://doi.org/10.1145/3355089.3356500 [34] Hyeongwoo Kim, Pablo Garrido, Ayush Tewari, Weipeng Xu, Justus Thies, Matthias Niessner, Patrick P´erez, Christian Richardt, Michael Zollh¨ofer, and Christian Theobalt. 2018. Deep Video Portraits. 37, 4, ስህተት 163 (2018), 14 ገጾች. ACM Transactions on Graphics https://doi.org/ 10.1145/3197517.3201283 [35] Pavel Korshunov and S´ebastien Marcel. 2019. Vulnerability Assessment and Detection of Deepfake Videos. In . IEEE, 1–6. Proceedings of the 2019 International Conference on Biometrics https://doi.org/10.1109/ICB45273.2019.8987375 [36] Patrick Kwon, Jaeseong You, Gyuhyeon Nam, Sungwoo Park, and Gyeongsu Chae. 2021. KoDF: A Large-scale Korean DeepFake Detection Dataset. In . IEEE, 10724–10733. Proceedings of the 2021 IEEE/CVF International Conference on Computer Vision https://doi.org/10.1109/ICCV 48922.2021.01057 [37] Lingzhi Li, Jianmin Bao, Hao Yang, Dong Chen, እና Fang Wen. 2020. FaceShifter: Towards High Fidelity And Occlusion Aware Face Swapping. arXiv:1912.13457. https://arxiv.org/abs/1912.13457 [38] Xurong Li, Shouling Ji, Chunming Wu, Zhenguang Liu, Shuiguang Deng, Peng Cheng, Min Yang, and Xiangwei Kong. 2021. Survey on Deepfakes and Detection Techniques. 32, 2 (2021), 496-518 ሶፍትዌር Journal http://www.jos.org.cn/1000-9825/6140.htm [39] Yuezun Li, Ming-Ching Chang, and Siwei Lyu. 2018. In Ictu Oculi: Exposing AI Created Fake Videos by Detecting Eye Blinking. In . IEEE, 1–7. Proceedings of the 2018 IEEE International Workshop on Information Forensics and Security https://doi.org/10.1109/WIFS.2018.8630787 [40] Yuezun Li, Xin Yang, Pu Sun, Honggang Qi, and Siwei Lyu. 2020. Celeb-DF: A Large-Scale Challenging Dataset for DeepFake Forensics. In . IEEE, 3204–3213. የ 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (የኮምፒውተር ቪዛ እና ሞዴሎች ማረጋገጫ) https://doi.org/10.1109/CVPR42600. 2020.00327 [41] Ruigang Liang, Peizhuo Lv, Yue Zhao, Peng Chen, Hao Xing, Yingjun Zhang, Jizhong Han, Ran He, Xianfeng Zhao, Ming Li, እና Kai Chen. 5, 2 (2020), 1–17. Journal of Cyber Security http://jcs.iie.ac.cn/xxaqxb/ch/re አግኝተዋል / አግኝተዋል abstract.aspx?file no=20200202&flag=1 [42] Steven R. Livingstone and Frank A. Russo. 2018. The Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS): A Dynamic, Multimodal Set of Facial and Vocal Expressions in North American English. 13, 5 (2018), 35 pages. PloS one [43] Chen-Chou Lo, Szu-Wei Fu, Wen-Chin Huang, Xin Wang, Junichi Yamagishi, Yu Tsao, and Hsin-Min Wang. 2021. MOSNet: Deep Learning based Objective Assessment for Voice Conversion. arXiv:1904.08352. https://arxiv.org/pdf/1904.08352.pdf [44] Jaime Lorenzo-Trueba, Junichi Yamagishi, Tomoki Toda, Daisuke Saito, Fernando Villavicencio, Tomi Kinnunen, እና Zhenhua Ling. 2018. The Voice Conversion Challenge 2018: Parallel እና Nonparallel Methods ልማት ለማሻሻል. International Speech Communication Association, 195-202 በዓለም አቀፍ ጓደኛ ጓደኛ ጓደኛ Proceedings of the Odyssey 2018 The Speaker and Language Recognition Workshop https://doi.org/10.21437/Odyssey.2018-28 [45] Siwei Lyu. 2020. Deepfake Detection: Current Challenges and Next Steps. In . IEEE, 6 pages. የ 2020 IEEE International Conference on Multimedia Expo Workshops (የ 2020 IEEE International Conference on Multimedia Expo Workshops) (የ 2020 IEEE International Conference on Multimedia Expo Workshops) (የ 2020 IEEE International Conference on Multimedia Expo Workshops) (የ 2020 IEEE International Conference on Multimedia Expo Workshops) https://doi.org/10.1109/ICMEW46912.2020.9105991 [46] Yisroel Mirsky and Wenke Lee. 2021. The Creation and Detection of Deepfakes: A Survey. 54, 1, Artikel 7 (2021), 41 ገጾች. ACM Computing Survey https://doi.org/10.1145/3425780 [47] Gautham J. Mysore. 2015. Can we Automatically Transform Speech Recorded on Common Consumer Devices in Real-World Environments into Professional Production Quality Speech?—A Dataset, Insights, and Challenges. 22, 8 (2015), 1006–1010. IEEE Signal Processing Letters https://doi.org/10.1109/LSP.2014.2379648 [48] Jo˜ao C. Neves, Ruben Tolosana, Ruben Vera-Rodriguez, Vasco Lopes, Hugo Proen¸ca, and Julian Fierrez. 2020. GANprintR: Improved Fakes and Evaluation of the State of the Art in Face Manipulation Detection. 14, 5 (2020), 1038–1048. የ IEEE Journal of Selected Topics in Signal Processing (አይኤይኤይኤይአይአይአይአይ) https://doi.org/10.1109/JSTSP.2020.3007250 [49] NIST Media Forensics Challenge Team. 2021. Open Media Forensics Challenge 2020-2021 ግምገማ ግምገማ ፕሮግራም. https://mig.nist.gov/MFC/Web/EvalPlan2020/OpenMFC2020EvaluationPlan.pdf [50] Aaron van den Oord, Sander Dieleman, Heiga Zen, Karen Simonyan, Oriol Vinyals, Alex Graves, Nal Kalchbrenner, Andrew Senior, እና Koray Kavukcuoglu. https://doi.org/10.48550/ARXIV.1609.03499 [51] Debajyoti Pal እና Tuul Triyason. 2018. የቪዲዮ አገልግሎቶች የሙከራ ጥራት ግምገማ ለማግኘት መደበኛ መተግበሪያዎች ግምገማ: አንድ ITU ግምገማ. 2018, ጽሑፍ 1391724 (2018), 25 ገጾች. International Journal of Digital Multimedia Broadcasting https://doi.org/10.1155/20 18/1391724 [52] Bo Peng, Hongxing Fan, Wei Wang, Jing Dong, Yuezun Li, Siwei Lyu, Qi Li, Zhenan Sun, Han Chen, Baoying Chen, Yanjie Hu, Shenghai Luo, Junrui Huang, Yutong Yao, Boyuan Liu, Hefei Ling, Guosheng Zhang, Zhiliang Xu, Changtao Miao, Changlei Lu, Shan He, Xiaoyan Wu, እና Wanyi Zhuang. https: //arxiv.org/abs/2106.01217 [53] Ivan Perov, Daiheng Gao, Nikolay Chervoniy, Kunlin Liu, Sugasa Marangonda, Chris Um´e, Mr. Dpfks, Carl Shift Facenheim, Luis RP, Jian Jiang, Sheng Zhang, Pingyu Wu, Bo Zhou, and Weiming Zhang. 2020. DeepFaceLab: Integrated, Flexible and Extensible Face-swapping Framework. https://doi.org/10.48550/ARXIV.2005.05535 [54] Stanislav Pidhorskyi, Дональд A. Adjeroh, እና Gianfranco Doretto. 2020. ተደጋጋጋሚ የመጨረሻ Au-to-coders. በ በ 10 ገጾች. Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition https://doi.org/10.1109/CVPR42600.2020.01411 [55] Andreas R ̈ossler, Davide Cozzolino, Luisa Verdoliva, Christian Riess, Justus Thies, እና Matthias Nießner. 2018. FaceForensics: በዓለም አቀፍ ደረጃ የቪዲዮ ውሂብ ስብስቦች. https://doi.org/10.48550/ARXIV.1803.09179 [56] Andreas R ̈ossler, Davide Cozzolino, Luisa Verdoliva, Christian Riess, Justus Thies, እና Matthias Nießner. 2019 የ FaceForensics++: የተመሠረተ የጽሑፎች ማረጋገጫ ለማግኘት ማወቅ. በ አዲሱ, 1 - 11 Proceedings of the 2019 International Conference on Computer Vision https://doi.org/10.110 9 / ICCV.2019.00009 ላይ [57] Anindya Sen. 2021. የአትክልት እና የቴክኒክ ኢንተርኔት: እንዴት የ "ዲፕፌካስ" የሙዚቃ ልምድ ለመፍጠር መረዳት ይችላሉ! https://medium.com/art-world-zen/art-and-artificia የኮምፒዩተር የኮምፒውተር የኮምፒውተር የኮምፒውተር የኮምፒውተር የኮምፒውተር የኮምፒውተር የኮምፒውተር የኮምፒውተር የኮምፒውተር የኮምፒውተር የኮምፒውተር የኮምፒውተር የኮምፒውተር የኮምፒውተር የኮምፒውተር የኮምፒውተር የኮምፒውተር የኮምፒውተር የ 7D29C [58] Kou Tanaka, Hirokazu Kameoka, Takuhiro Kaneko, and Nobukatsu Hojo. 2019. WaveCycleGAN2: Time-domain Neural Post-filter for Speech Waveform Generation. https://doi.org/10.48550/A የኦሪጂናል ስቴትስ 1904.02892 [59] Jianhua Tao, Ruibo Fu, Jiangyan Yi, Chenglong Wang, and Tao Wang. 2020. Development and Challenge of Speech Forgery and Detection. 5, 2 (2020), 28-38 Journal of Cyber Security http://jcs.iie.ac.cn/xxaqxb/ch/reader/view abstract.aspx?file no=20200204&flag=1 [60] Tencent. 2020 የፕላስቲክ ኢንተርኔት ቫይክሌት. https://tech.sina.com.cn/roll/2020-07-14/doc-iivhvpwx5201226.shtml [61] Justus Thies, Michael Zollh¨ofe, and Matthias Niessner. 2019. Deferred Neural Rendering: Image Synthesis using Neural Textures. 38, Article 66 (2019), 12 pages. Issue 4. ግራፊክስ ላይ ACM Transactions https://doi.org/10.1145/3306346.3323035 [62] Tomoki Toda, Ling-Hui Chen, Daisuke Saito, Fernando Villavicencio, Mirjam Wester, Zhizheng Wu, and Junichi Yamagishi. 2016. The Voice Conversion Challenge 2016. In International Speech Communication Association, 1632-1636. የ 2016 Interspeech ሂደቶች https://doi.org/10.21437/Interspeech.2016-1066 [63] Massimiliano Todisco, Xin Wang, Ville Vestman, Md Sahidullah, Hector Delgado, Andreas Nautsch, Junichi Yamagishi, Nicholas Evans, Tomi Kinnunen, and Kong Aik Lee. 2019. ASVspoof 2019: Future Horizons in Spoofed and Fake Audio Detection. arXiv:1904.05441. https://arxiv.org/ ግምገማዎች / ግምገማዎች [64] Ruben Tolosana, Ruben Vera-Rodriguez, Julian Fierrez, Aythami Morales, እና Javier Ortega-Garcia. 2020. Deepfakes እና ከዚያ በላይ: የጽሑፍ መተግበሪያ እና የጽሑፍ ማረጋገጫ ጥናት. 64 (2020), 131-148 የምስክር ወረቀት https://doi.org/10.1016/j.inffus.2020.06.014 [65] Xin Tong, Luona Wang, Xiaoqin Pan, እና Jingya Wang. በ 2020 ላይ የ Deepfake: The Sword of Damocles በ AI. . IEEE, 265–273. በ 2020 የኮምፒውተር ቪዛ, ፎቶ እና ጥልቅ መግዛት ዓለም አቀፍ ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር https://doi.org/10.1109/CVIDL51233.2020.00-88 [66] Luisa Verdoliva. 2020 Media Forensics እና DeepFakes: አንድ ግምገማ. 14, 5 (2020), 910 – 932 የ IEEE Journal of Selected Topics in Signal Processing (አይኤይኤይኤይአይአይአይአይ) https://doi.org/10.1109/JSTSP.2020.3002101 [67] Xin Wang, Junichi Yamagishi, Massimiliano Todisco, H ́ector Delgado, Andreas Nautsch, Nicholas Evans, Md Sahidullah, Ville Vestman, Tomi Kinnunen, Kong Aik Lee, Lauri Juvela, Paavo Alku, Yu-Huai Peng, Hsin-Te Hwang, Yu Tsao, Hsin-Min Wang, S ́ebastien Le Maguer, Markus Becker, Fergus Henderson, Rob Clark, Yu Zhang, Quan Wang, Ye Jia, Kai Onuma, Koji Mushika, Takashi Kaneda, Yuan Jiang, Li-Juan Liu, Yi-Chiao Wu, Wen-Chin Huang, Tomoki Toda, Kou Tanaka, Hirazu Kameoka, Ingmar Steiner, Driss Matrouf, Jean-Francois Bonastre 64 (2020), 27 pages. የኮምፒውተር ስሜት & ቋንቋ https://doi.org/10.1016/j.csl.2020.101114 [68] Mirjam Wester, Zhizheng Wu, እና Junichi Yamagishi. 2016 የሙዚቃ መተግበሪያ ልምድ ልምድ 2016 ግምገማ ውጤቶች. በ International Speech Communication Association, 1637-1641 በዩናይትድ ስቴትስ. Interspeech 2016 ኮምፒዩተር ኮምፒዩተር https://doi.org/10.21437/Interspeech.201 6-1331 አግኙን [69] Junichi Yamagishi, Massimiliano Todisco, Md Sahidullah, H´ector Delgado, Xin Wang, Nicholas Evans, Tomi Kinnunen, Kong Aik Lee, Ville Vestman, and Andreas Nautsch. 2019. ASVspoof 2019: Automatic Speaker Verification Spoofing and Countermeasures Challenge Evaluation Plan. https://www.asvspoof.org/asvspoof2019/asvspoof2019 evaluation plan.pdf [70] Zhao Yi, Wen-Chin Huang, Xiaohai Tian, Junichi Yamagishi, Rohan Kumar Das, Tomi Kinnunen, Zhen-Hua Ling, እና Tomoki Toda. 2020 የሙዚቃ መተግበሪያ ልምድ 2020 – የሙዚቃ መተግበሪያ እና የሙዚቃ መተግበሪያ ልምድ –. . International Speech Communication Association, 80–98. የ Blizzard Challenge እና የ Voice Conversion Challenge 2020 ተመራማሪዎች https://doi.org/10.21437/VCC BC.2020-14 [71] Mohammad A. Younus እና Taha M. Hasan. 2020 የ Deepfake ቪዲዮ ማረጋገጫ ቴክኖሎጂዎች ጥቅል. በ የ IEEE, 115-120 የ 2020 የ 6 ኛው ዓለም አቀፍ ኢንጂነሪንግ ኮምፒዩተር https://doi.org/10.1109/IEC49899.2020.9122916 [72] Guangtao Zhai እና Xiongkuo ደቂቃ 2020. Perceptual Image Quality Assessment: A Survey. 63, Article 211301 (2020), 52 pages. የቴክኖሎጂ ሳይንስ ቻይና https://doi.org/10.1007/s11432-019-2757-1 [73] Teng Zhang, Lirui Deng, Liang Zhang, እና Xianglei Dang. 2020 ዱቄት አጠቃቀም: የዱቄት አጠቃቀም ላይ ግምገማ. በ 67 - 70 ውስጥ. የ 2020 IEEE 3rd International Conference on Computer and Communication Engineering Technology (የኮምፒውተር እና የኮምፒውተር እና የኮምፒውተር ኢንጂነሪንግ ቴክኖሎጂ) https://doi.org/10.1109/CCET5090 1.2020.9213159 አግኝቷል [74] Yuxuan Zhang, Huan Ling, Jun Gao, Kangxue Yin, Jean-Francois Lafleche, Adela Barriuso, Antonio Torralba, and Sanja Fidler. 2021. DatasetGAN: Efficient Labeled Data Factory with Minimal Human Effort. In . IEEE, 10140–10150. የ 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (የኮምፒውተር ቪዛ እና ሞዴሎች ማረጋገጫ) https://doi.org/10.1109/CVPR46437.2021.01001 [75] ![](file:///C:/Users/user/AppData/Local/Temp/msohtmlclip1/01/clip_image051.gif)Yuanhan Zhang, ZhenFei Yin, Yidong Li, Guojun Yin, Junjie Yan, Jing Shao, and Ziwei Liu. 2020. CelebA-Spoof: Large-Scale Face Anti-spoofing Dataset with Rich Annotations. In ስፕሪንግር 70-85 Proceedings of the 2020 European Conference on Computer Vision https://doi.org/10.1007/97 8-3-030-58610-2 5 [76] Yuanhan Zhang, Zhenfei Yin, Jing Shao, Ziwei Liu, Shuo Yang, Yuanjun Xiong, Wei Xia, Yan Xu, Man Luo, Jian Liu, Jianshu Li, Zhijun Chen, Mingyu Guo, Hui Li, Junfu Liu, Pengfei Gao, Tianqi Hong, Hao Han, Shijie Liu, Xinhua Chen, Di Qiu, Cheng Zhen, Dashuang Liang, Yufeng Jin, እና Zhanlong Hao. 2021. የፀሐይ-Spoof Challenge 2020 ላይ የፊት መከላከያ: ዘዴዎች እና ውጤቶች. https://arxiv.org/pdf/2102.12642.pdf [77] Bo Zhao, Shaozeng Zhang, Chunxue Xu, Yifan Sun, እና Chengbin Deng. 2021. ጥቁር የጂኦግራፊያ? 48, 4 (2021), 338–352. የኮርቶግራፊያ እና የጂኦግራፊያ መረጃ ሳይንስ https://doi.org/10.1080/15230406.2021.1910075 [78] Peng Zhou, Xintong Han, Vlad I. Morariu, እና Larry S. Davis. 2017 ለ Tampered Face Detection ለ ሁለት-ስሬም የኒውሮል አውታረ መረብ. በ በ IEEE, 1831-1839 ውስጥ. የ 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (የኮምፒውተር ቪዛ እና ሞዴሎች ማረጋገጫ ላይ የ 2017 IEEE Conference (የኮምፒውተር ቪዛ እና ሞዴሎች ማረጋገጫ) https://doi.org/10.1109/CVPRW.2017.229 [79] Tianfei Zhou, Wenguan Wang, Zhiyuan Liang, እና Jianbing Shen. 2021. በይፋ ውስጥ የኮርፖሬሽን. በ IEEE, 5774-5784 ላይ. የ 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (የኮምፒውተር ቪዛ እና ሞዴሎች ማረጋገጫ) https://doi.org/10.1109/CVPR46437.2021.00572 [80] Jun-Yan Zhu, Taesung Park, Phillip Isola, እና Alexei A. Efros. 2017 በሲክሌት-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ-የተካከለ በ IEEE, 2242-2251 ላይ. የ 2017 IEEE International Conference on Computer Vision (የኮምፒውተር ቪዛ (የኮምፒውተር ቪዛ) ዓለም አቀፍ ኮምፒውተር ኮምፒውተር ኮምፒውተር ኮምፒውተር https://doi.org/10.1109/ICCV.2 017.244 [81] Bojia Zi, Minghao Chang, Jingjing Chen, Xingjun Ma, እና Yu-Gang Jiang. 2020 WildDeepfake: የ Deepfake ማረጋገጫ ለማግኘት አስቸጋሪ እውነተኛ ዓለም ውሂብ ስብስቦች. በ የ ACM, 2382–2390 የ 28th ACM International Conference on Multimedia (ኤስ.ኤስ.ኤስ.ኤስ.ኤስ.ኤስ.ኤስ.ኤስ.ኤስ.ኤስ.ኤስ.ኤስ.ኤስ.ኤስ.ኤስ.ኤስ.ኤስ.ኤስ.ኤስ.) https://doi.org/10.1145/339417 1.3413769 This paper is under CC by 4.0 Deed (Attribution 4.0 International) license. available on arxiv ይህ ገጽ በ CC by 4.0 Deed (Attribution 4.0 International) አግኝቷል. አግኝቷል Archive