The innovation of Real-time Facial Analysis System is designed to employ like neural network, using lots of node and complex layer to train a computer to learn and recognize the human facial features from images.
By applying the state-of-the-art machine learning method with CelebA database, the technology can recognize exceed 40 kinds of facial attributes, including gender, hair color, smiling, makeup, wearing necklace and so on.
The innovation has proven to verify human facial expression at different angles and illumination; this is different from the shallow face recognition methods which are limited on frontal faces and suffered from the variance of illumination and expression.
Global
