Compressed Deep Convolution Neural Network for Face Recognition

Deep convolution neural network (CNN) has achieved a great success on face recognition techniques. But most of CNN models tend to be much deeper, which are at the expenses of high consumption of computation and storage. So, it is hard for these deep CNNs applied to mobile equipments because of poor computational and memory resources. To alleviate this issue, this paper optimizes a lightened baseline CNN model by adopting an additional contrastive loss to learn more discriminative features. To further reduce the number of parameters, a pruning strategy is tried to compress our model, which slightly improves accuracy on the LFW dataset with the compression ratio of 0.7. Finally, experimental result shows that the proposed method achieve state-of-the-art results with much smaller size and fewer training data.