Face recognition method and device based on face attribute perception loss, and electronic device

Abstract

The embodiment of the invention provides a face recognition method and device based on face attribute perception loss, and an electronic device. The method comprises: carrying out the scale normalization processing of a face image on the RGB modal data and Depth modal data of a face to be recognized, and carrying out the channel stacking fusion of the normalized face RGB modal data and the normalized face Depth modal data; inputting the face RGB modal data and the face Depth modal data which are subjected to stacking fusion into a deep convolutional network, and extracting high-dimensional face features of a face to be identified; inputting the high-dimensional face features into a face attribute perception loss and classification loss combination layer to realize face attribute-based feature clustering; and based on the feature clustering, realizing recognition of the to-be-recognized face by calculating similarity scores with different face images in the face image library. Accordingto the embodiment of the invention, the feature expression of the face can be more effectively optimized, so that the face recognition can be more accurately carried out.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of face recognition, and more specifically, relate to a face recognition method, device and electronic equipment based on perceptual loss of face attributes. Background technique [0002] In recent years, with the rapid development of deep learning technology, face recognition, as one of the hot issues in the field of computer vision, has made great progress. At present, most of the mainstream face recognition methods first use deep networks to map face samples to high-dimensional feature spaces to obtain discriminative feature expressions, and then classify faces by calculating feature distances. [0003] For example, the contrastive loss function (contrastive loss), the triplet loss function (triplet loss), and a series of softmax variant losses represented by center loss, all use the method of metric learning to reduce the distance of intra-class features, while increasing the class ...

AI Technical Summary

Problems solved by technology

However, these methods do not take into account the potential semantic expression of face samples, causing their feature distribution in high-dimensional space to tend to be evenly distributed.

That is to say, when the traditional deep learning loss function is used for face recognition, the similarity of the features of different types of samples cannot represent the potential semantic similarity of the samples, so it is not conducive to more accurate face recognition.

Images