Facial image processing with convolutional neural networks

We present a generic machine-learning technique based on Convolutional Neural Networks and describe how it can be successfully applied to different facial processing applications like face detection, facial feature detection, face alignment, gender classification and face recognition. We experimentally show that the proposed approach outperforms state-of-the-art techniques based on Adaboost, SVMs or AAMs, as it conjointly teams feature extraction and classification parameters, making it very robust to noise, lighting variations and partial occlusions while being naturally and efficiently implemented on embedded platforms.