Too Long; Didn't Read
Build a Smart Baby Monitor Using a RaspberryPi and Tensorflow. It should run on anything as simple and cheap as a cheap RaspberryPi with a cheap USB microphone. The same exact procedure can be used to use sound to recognize a baby’s cries, as long as they’re long enough to long enough over the background noise. We’ll have to record enough audio samples where the baby doesn’t cry and where the cry is in the background. Note: In this example I’m using the RaspberryPi to train the new model in how to detect cries in any type of sound.