I would be interested to know how well you might think that openMV might work for emotion recognition. Training of the prediction model/would be done on an external device/pc. Im thinking here whats possible.
Im thinking something like this project
Im after feasibility for realtime with video.
All the best
Yes I think this a perfect application for OpenMV CNN, especially that we have a fast face detector (see the smile detection example). We’re going to release the updated fw, ide and examples very soon also we’re posting a training tutorial, so you could try to train it yourself if you like. If the network works well we can add it to the repo.
Actually we just posted the tutorial, here’s the link