This script will activate your webcam, read your facial expression based on a pretrained model (.h5) and make the screen flash if you look angry for too long while working ;) It gets users to relax their facial expressions since you get quite aware of your expression after the 5th flash. Since the neurological feedback generated by tensing/relaxing facial muscles this does actually change your mood a bit. If you have a very tense facial expression while working try it out.
I didnt include the dataset of images for facial expressions as i wanted to keep size down to a minimum.
Screen_Flash_When_Angry.py Use this script to activate the camera and starts the face recognition. ( screen flashes )
Doggy_Pic_When_Angry.py Is another version of the Screen_Flash_When_Angry.py script Use this script to activate the camera and starts the face recognition. ( doggo pic pops up )
Emotion_rec_model.h5 Is the trained model with all weights and architecture saved
Classification_CNN.py Was used to train the model. accuracy is around 65% but as all simple emotion_rec models it depends on light/face. I dont like the model
haarcascade_frontalface_default.xml Is the cascade used for face recognition in openCV model
requirements.txt Defines all packages needed
Please do so but state / original github account or as contributor