Face recognition method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention relates to the field of artificial intelligence, and discloses a face recognition method and device, electronic equipment and a storage medium. The method comprises thesteps of inputting a to-be-recognized face image into a pre-trained convolutional neural network model to obtain a recognition result output by the convolutional neural network model, wherein the convolutional neural network comprises an affine estimation branch network and a convolutional integral branch network, the affine estimation branch network is used for face detection and face alignment,the convolution branch network is used for carrying out identity recognition on the face image. An affine estimation branch network is added to a convolutional neural network model and used for achieving face detection and face alignment, face detection, face alignment and face recognition are further completed in the convolutional neural network model at the same time, and therefore end-to-end face recognition is achieved.

Description

technical field [0001] Embodiments of the present invention relate to the field of artificial intelligence, in particular to the field of computer vision. Background technique [0002] Traditional face recognition techniques can basically be classified into three categories, namely: geometric feature-based methods, template-based methods, and model-based methods. In the past few years, the introduction of convolutional neural networks (CNN) has greatly improved The performance of existing computer vision tasks has made it gradually become a mainstream method, including facial recognition and verification. Instead of constructing a classification model by manually labeling features, deep learning successfully improves the robustness of facial features through data-driven. Therefore, a well-trained CNN network can better handle the pose, occlusion and illumination changes of face images. [0003] However, large pose changes in real-life scenarios are still a major challenge ...

AI Technical Summary

Problems solved by technology

[0004] The inventors found that there are at least the following problems in the prior art: before face feature extraction, it is necessary to perform face detection according to the prior knowledge of the face structure and then align the face regions, resulting in the failure of the end-to-end face recognition model. Training and design are difficult because face detection, face alignment and face recognition are performed separately and independently

Images