Human body behavior recognition method and system based on motion history images

Abstract

The invention discloses a human body behavior identification method and system based on a motion history image, and the method comprises the steps of extracting a video image frame based on a human body motion video image, and dividing the attribute of each pixel point into a foreground pixel point and a background pixel point through an inter-frame difference method; obtaining a pixel point attribute change sequence at the same position; improving the motion history map MHI to obtain a video image grey-scale map under each change mode; performing feature extraction by taking the grey-scale map under each change mode as a feature image; inputting into a preset classifier model for action behavior identification. According to the invention, an improved motion history map MHI is adopted, theMHI is extracted by using the MHI, the spatial domain information in the MHI is extracted by using 2DHaar wavelet transform, and the time domain information in the MHI is extracted by using the statistical histogram of the MHI, so that the operation complexity is reduced, the feature contains richer motion information, and an identification algorithm with a simple classification process and a high speed is realized.

Description

technical field [0001] The invention relates to the technical field of behavior recognition, in particular to a human behavior recognition method and system based on motion history images. Background technique [0002] Behavior recognition is becoming more and more important in current computer vision applications, especially behavior classification. In order to understand what happened in the scene, we must perform behavior recognition. Behavior recognition is also the basis of human-computer interaction. Human behavior has always been a difficult problem in video surveillance, video classification, and social scenes. Due to the great variability of human behavior, template-based recognition methods have great limitations. Due to the structural characteristics of the human body, we need multi-dimensional Space to describe, and due to the non-rigidity of the clothes, the behavior recognition problem is further complicated. Therefore, we need a nonlinear model for modeling an...

AI Technical Summary

Problems solved by technology

[0002] Behavior recognition is becoming more and more important in current computer vision applications, especially behavior classification. In order to understand what happened in the scene, we must perform behavior recognition. Behavior recognition is also the basis of human-computer interaction. Human behavior has always been a difficult problem in video surveillance, video classification, and social scenes. Due to the great variability of human behavior, template-based recognition methods have great limitations. Due to the structural characteristics of the human body, we need multi-dimensional Space to describe, and due to the non-rigidity of the clothes, the behavior recognition problem is further complicated. Therefore, we need a nonlinear model for modeling and optimization, and nonlinearity is particularly computationally complex.

[0003] In behavior recognition technology, the selection and expression of features is very important, including features with rich information and strong discrimination, which can strengthen the recognition effect, and is an important part of the whole process that cannot be underestimated. On the one hand, the feature extraction process requires It consumes a lot of storage space and computing resources to extract motion features. On the other hand, human behavior itself is dynamically generated by the change of human posture along with time. Actions cause wrong recognition, because many types of actions are in the process of changing. If you look at it regardless of time, there are many similarities in shape, and even the content of certain frames is the same. In the prior art, the motion history map MHI Although the order of actions is reflected by the gray value of pixels, it can only fully reflect the order of actions in one-way motion actions. For round-trip motion, very close actions appearing in different process sequences, such as sitting down With the paired action of standing up, the motion history map MHI cannot clearly express the information of time sequence, that is to say, the motion history map MHI in the prior art cannot represent the actual motion. The accuracy of the feature vector extracted by the historical graph MHI is not high

Images