Gait recognition method and system combining subspace learning and tensor neural network

Abstract

The invention discloses a gait recognition method and system combining subspace learning and tensor neural network, belonging to the field of intelligent recognition. The method includes: acquiring gait data, obtaining a gait data set, and processing to obtain a set of silhouette images, and further obtaining a gait energy image; taking 80% of the silhouette images as a training set for dimension reduction processing, and the remaining 20% ​​as a test set The data is used to test the training results, and then the gait energy map and the data after dimensionality reduction are used for feature extraction through the tensor neural network module, and then the support vector machine is used as a classifier for classification, and finally the results of the training set and the test set are compared. The result of identifying and identifying the identity of pedestrians is obtained by comparison. The invention has a simple implementation method and low hardware cost, can automatically detect the identity authority of a person in a specific place and identify the identity of a fake person, and effectively improve the safety protection of the monitoring place and the identity identification in various situations.

Description

technical field [0001] The invention relates to the field of intelligent recognition, in particular to a gait recognition method and system combining subspace learning and tensor neural network. Background technique [0002] In order to better protect the property safety and information security of individuals, families and enterprises, many places are equipped with safety protection measures. Traditional security mainly relies on manpower, such as setting up guards for registration, organizing personnel to patrol, etc. Since this method cannot guarantee that there are guards in each area at any time, it is prone to security loopholes. In recent years, with the rapid development of computer, network, image processing and transmission technology, gait recognition technology has shown its value more and more. Through the processing of video materials, gait recognition image data can be obtained, which can avoid fingerprint forgery, mask anti-face recognition, etc. Due to the ...

AI Technical Summary

Problems solved by technology

[0003] However, most of the current personnel identification is essentially a human monitoring method, that is, by videotaping the environment in the monitoring area and relying on the naked eyes of the monitoring personnel to find abnormalities, it is impossible to accurately detect people in certain situations , cannot avoid counterfeiting measures such as fingerprint imitation and wearing a mask,

This artificial monitoring method has high cost, poor flexibility, large limitations, high requirements for human body activities and wear, and cannot meet the real needs of users, that is, some precise identification needs.

Images