Similarity mining method of similar behaviors between multiple views and behavior recognition method

Abstract

The invention discloses a similarity mining method of similar behaviors between multiple views and a behavior recognition method. The behavior recognition method is based on the similarity mining method. The similarity mining method comprises calculating a Hessian matrix response of every pixel in every frame of gray images of video streaming of different views of the similar behaviors frame by frame; determining feature points according to the matrix responses; building space-time feature descriptors of the feature points and obtaining posture description vectors after dimensionality reduction; calculating a similarity recurrence plot matrix with the posture description vectors as low-level features; and obtaining recursiveness descriptors according to the recurrence plot matrix and clustering the recursiveness descriptors. According to the similarity mining method of the similar behaviors between the multiple views and the behavior recognition method, posture modeling is achieved through a space-time feature point method and recursive similarity between the multiple views is mined through a recurrence plot, so that the problem that background segmentation dependency is easy to cause distortion of human body postures is solved and the accuracy of similarity mining and the robustness of behavior recognition are improved.

Description

technical field [0001] The invention belongs to the technical field of view image processing, and in particular relates to a similarity mining method among multiple views of the same behavior and a behavior recognition method based on the similarity mining method. Background technique [0002] Visual recognition and understanding human behavior is still an active research area in computer vision, and solving this problem has potential for many applications, such as video retrieval, large structured video archives, video surveillance, human-computer interaction, gesture recognition, and video editing Wait. Behavioral understanding has been approached in many different ways in recent years. Since the human body pose profile has distinct semantics, which can clearly express the behavior category, the time transformation relationship of the human body pose profile is often used as an effective behavior description. In the process of behavior recognition, the same behavior will...

AI Technical Summary

Problems solved by technology

This method has the following disadvantages: (a) Extracting the behavior trajectories of the head, hands and feet of key human body parts will be disturbed by factors such as occlusion, illumination changes, and inaccurate segmentation of these parts, which will lead to behavioral problems in practical applications. It is not robust, and the probability of error increases; (b) Due to the interference in the actual application, the position of the "center of mass" of the head, hands and feet of the moving part will be wrong, resulting in a deviation in the behavior trajectory, which makes the calculation of the cross-ratio of the trajectory point There will also be large errors, which will affect the recognition accuracy

The disadvantages of this recognition method are: (a) the method of extracting binary human body contours in practical applications will be disturbed by factors such as occlusion and illumination changes, which will make the human body contours distorted; When the following binary human body contour information is used for sculpting and reconstruction of the human body's 3D posture, it is necessary to calculate the corresponding matching points between the three views, which not only requires a large amount of calculation, but also easily causes matching errors due to the lack of outstanding features of the corresponding points.

Images