TLDR
With the spread of COVID-19 wearing face masks became obligatory. Apple’s FaceID struggled to recognize faces with masks. We created such data augmentations that transform the first training dataset into that of the face with a medical mask. The database of medical masks is stored in. JSON. It includes calculated parameters of rotation. The solution can be found in the solution repository via this link. You can replicate the whole process with this colab notebook and even prepare your own dataset with a provided pipeline.via the TL;DR App
no story
Written by broutonlab | BroutonLab helps businesses optimize costs and drive revenues with bespoke AI Solution
