Using deep learning to detect smiling faces.
To detect smiling faces on pre-recorded video or live cam feed. Using Keras and OpenCV
-> numpy -> matplotlib -> openCV -> sklearn -> imutils -> TensorFlow/Theano -> Keras -> h5py
- Gather a dataset of people smiling and not smiling people which is tighly cropped around the face(to remove noise).
- Train the Convolutional Neural Network using
train_model.py
and your custom data set(the network is designed to handle class imbalance). - Run the
detect_smile.py
to start a live video using your webcam, or you can also pass an optional argument to enter the path to a pre-recorded video.