Multi-modality recognition of human face and ear based on deep learning

Abstract

Deep learning is used for learning better features through constructing a model of machine learning, which contains a lot of hidden layers. At present, deep learning is widely used in face recognition. But faces of different ages, facial expressions and other changes will affect the result of face recognition. The human ear is not affected by different ages, facial expressions, makeups and so on, which makes the ear recognition has certain research value and application prospects. But ears are more likely to be blocked than faces. The multi-modality identification of multiple biometric features could make the identification accuracy higher, as a result, human ears are used in the face recognition system to improve the accuracy of the identification system. In the paper, Faster R-CNN is used to detect human faces and ears firstly. Then CNNs are used to train our own nets and recognize people using their faces and ears respectively in the images. Finally, Bayesian decision fusion method is used to fuse the recognition results of the human face and ear to improve the recognition rate.