Human motion recognition method based on depth movement trail

Abstract

The invention discloses a human motion recognition method based on a depth movement trail. The human motion robustness based on depth information is described and recognized. The method specifically includes the following steps of (1) video preprocessing, (2) establishment of multiple scale spaces, (3) space meshing and feature point extraction of each scale space, (4) feature point tracking, (5) description based on the depth movement trail, (6) feature normalization based on a 'word bag' method, and (7) motion recognition based on an SVM. The method has the advantages that human motions are described through the depth movement trail, robustness and effectiveness are quite good, and efficient human motion recognition is achieved.

Description

technical field [0001] The invention belongs to the technical field of computer vision and pattern recognition, and relates to a human action recognition method based on a deep motion trajectory, which is used for robust and efficient description of deep human action, thereby performing efficient human action recognition. Background technique [0002] Human action recognition is a very active research topic in the field of computer vision, and with the development of computers, it has been widely used in many fields, such as video surveillance, human-computer interaction and video analysis. With the release of Kinect by Microsoft, more and more researchers began to pay attention to the depth data in Kinect. Compared with RGB data, depth data has the following advantages: First, depth data can provide pure geometry and clear edges, so it is better than color and texture in RGB data in image segmentation, target recognition, and action recognition. Second, depth images are i...

AI Technical Summary

Problems solved by technology

Although some action recognition algorithms based on depth data have been proposed, due to the limitations of depth sensors, the obtained depth data has the following characteristics: 1) The pixel value jumps relatively large, especially in the edge area; 2) In the same position , the depth values ​​are the same, and the difference is small

It is precisely because of the above reasons that it is difficult to describe the depth motion data, and the performance of the existing motion recognition algorithms based on depth data needs to be improved.

Images