Human face recognition based on adaptive deep Convolution Neural Network

Convolution Neural Network among most of the methods for recognition has a more desirable recognition accuracy. This work introduces an adaptive learning algorithm with an adaptive learning rate, and the algorithm is applied into the human face recognition. It solves the problems in choosing the appropriate rate and improves the slow process when faced with big data. The algorithm is tested in the FERET datasets, and is compared with the traditional deep convolution neural network. The test results show that this algorithm do increase the speed of convergence and reduce the recognition errors.