Cross-visual-angle gait identification method based on tensor simultaneous discriminant analysis

Abstract

The invention provides a cross-visual-angle gait identification method based on tensor simultaneous discriminant analysis. The cross-visual-angle gait identification method comprises steps of constructing a gait characteristic expressed on the basis of Gabor, online model training and an offline test. The cross-visual-angle gait identification method not only adopts a coupling measurement study principle to weak a data heterogeneous problem under the cross-visual-angle, combines with the tensor gait characteristic which is expressed on the basis of the Gabor and the tensor discriminant analysis principle, improves classification performance of the gait and avoids a problem of a small sample because of lack of samples.

Description

technical field [0001] The invention relates to a cross-view gait recognition method based on tensor simultaneous discriminant analysis, which belongs to the field of pattern recognition and machine learning. technical background [0002] Gait is one of the most potential biological characteristics at a long distance. It has the advantages of non-contact acquisition, not easy to imitate camouflage, less affected by the environment, and occupies less memory. It has a wide range of applications in the fields of public safety, medical diagnosis, and case investigation. Application prospect and economic value. [0003] Under controllable conditions, the current research results have achieved good recognition results. However, gait recognition faces many difficulties, such as changes in time, clothing changes, walking surfaces, shoes and hats, backpacks and other carrying objects, and viewing angles. In particular, changes in pedestrian gait viewing angles will lead to observabl...

AI Technical Summary

Problems solved by technology

However, gait recognition faces many difficulties, such as changes in time, clothing changes, walking surfaces, shoes and hats, backpacks and other carrying objects, and viewing angles. In particular, changes in pedestrian gait viewing angles will lead to observable human gait images and registration samples. There are differences in the characteristics between them, which leads to a great reduction in the accuracy of gait recognition

[0004] In order to solve the problem of cross-view gait recognition, researchers have proposed a large number of cross-view gait recognition methods, which can be classified into four categories according to the characteristics of these methods: one is to find insensitivity features, Kale et al. through sagittal plane perspective projection Synthesize the gait at any angle with the gait image of a specific angle of view. When the angle of view of the sagittal plane and the image plane is very different, this method will cause serious self-occlusion, resulting in a significant decline in recognition performance; the second type is to use complex multiple Camera collaboration system collects 3D gait information to reconstruct gait samples from any viewing angle

In 2001, Shakhnarovich et al. proposed a standardization method for gait angle of view. In gait recognition, pedestrian motion information is updated synchronously through multiple cameras, and the typical angle of view is defined by the three-dimensional structure information of gait motion for gait recognition. This method is largely Relying on multi-camera cooperative operation under controllable conditions, and the cost is high and the operation is complicated; the third type is to use the perspective transformation model to learn the mapping relationship between gait samples under different perspectives, and convert the samples under the perspective to be tested to the registered samples perspective, and then feature extraction and sample identification

In daily life, there are often differences in perspective between the gait samples captured in the monitoring system and the registered gait samples, which greatly weakens the effectiveness of classic gait recognition algorithms.

Images