Video face recognition method based on recurrent neural network and video face recognition system based on recurrent neural network

Abstract

The invention provides a video face recognition method based on a recurrent neural network and a video face recognition system based on the recurrent neural network. The video face recognition methodbased on the recurrent neural network comprises the steps that an image frame is extracted from a video stream, the face target is tracked and the corresponding face sequence of the face target is acquired; the face sequence is preprocessed; the preprocessed face sequence is inputted to the recurrent neural network to be trained, and the weight value of each layer of the recurrent neural network is updated so as to obtain the trained recurrent neural network; the preprocessed face sequence is inputted to the trained recurrent neural network and the depth features of the face sequence are extracted; a face classifier is trained by using the depth features of the face sequence; the face features are extracted by using the preprocessed face sequence according to the trained face classifier; and the face features are compared with the feature template of a target library, and the face recognition information matched with the current face features in the target library is returned. According to the video face recognition method based on the recurrent neural network and the video face recognition system based on the recurrent neural network, the accuracy of video face recognition can beenhanced.

Description

technical field [0001] The invention relates to a video face recognition method and system, in particular to a video face recognition method and system based on a cyclic neural network. Background technique [0002] With the continuous upgrading of the network and the popularization of video shooting equipment, the shooting and dissemination of video in daily life has become very convenient, and the content of video format is also more and more widely loved by people. In the field of security, video surveillance has become an essential means of security, and monitoring equipment is widely distributed in every corner of the world. In the financial industry, banks and securities companies are also gradually collecting videos of account holders for identity verification. Therefore, the demand for identification through video is becoming more and more. [0003] In daily life, video shooting is usually performed by a handheld device. Since the position of the camera is in an un...

AI Technical Summary

Problems solved by technology

This recognition method is an extension of static face recognition technology to video face recognition. It does not have the characteristics of video, and the face-by-face comparison loses the relevant information between the face images in the face sequence.

In addition, it is difficult to judge what kind of face is more effective for comparison; using the location information of key points of the face and photo quality evaluation itself will introduce errors

When merging multiple comparison results, it is also difficult to weigh the results with large differences

Therefore, the recognition accuracy is also very unsatisfactory, and it is difficult to meet the needs of practical applications.

Images