Face dynamic recognition method based on 3D convolutional neural network and face dynamic recognition system based on 3D convolutional neural network

Abstract

The invention provides a face dynamic recognition method based on a 3D convolutional neural network and a face dynamic recognition system based on the 3D convolutional neural network. The face dynamicrecognition method based on the 3D convolutional 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 so as to acquire the face sequence meeting the preset standard; the preprocessed face sequence is inputted to the 3D convolutional neural network to be trained, and the weight value of each layer of the 3D convolutional neural network is updated so as to obtain the trained 3D convolutional neural network; the preprocessed face sequence is inputted to the trained 3D convolutional neural network and the face features of the face sequence are extracted; 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 face dynamic recognition method based on the 3D convolutional neural network and the face dynamic recognition system based on the 3D convolutional neural network, the face sequence is extracted from the video to be inputted to the 3D convolutional neural network to learn the face features in the video so that the accuracy of video face recognition can be enhanced.

Description

technical field [0001] The present invention relates to a face recognition method and system, in particular to a face dynamic recognition method and system based on a 3D convolutional neural network. Background technique [0002] At present, video surveillance equipment has been widely used and spread all over the public places in the city. Finding and tracking the target by identifying the identity of the person in the surveillance video has also become an effective way to improve work efficiency. Since monitoring devices may be distributed in various scenes, affected by factors such as light, angle, and device resolution, it is difficult to collect high-quality face images for identity verification. [0003] Since the target face often appears in multiple frames of video, it is a feasible way to make full use of multiple face images of the same person in the video to improve the recognition accuracy. Existing face recognition technologies mainly include the following two...

AI Technical Summary

Problems solved by technology

In these cases, it is difficult to select suitable face photos from the face sequence for comparison, and it is also difficult to combine the comparison results to obtain the final result.

Images