A Human Behavior Recognition Method Based on Depth Video Sequence

A deep video sequence and recognition method technology, applied in the field of computer pattern recognition, can solve the problems of reducing the recognition rate, misrepresenting coefficients, and not being able to more accurately represent the detailed information of local descriptors, so as to achieve strong dictionary expression ability and improve recognition rate Effect

CN104298974BActive Publication Date: 2018-03-09BEIJING UNIV OF TECH
1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Publication Date
2018-03-09

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention discloses a human behavior recognition method based on a depth video sequence. The method calculates the four-dimensional normal vectors of all pixels in the video sequence, and extracts pixels in different layers by constructing a time-space pyramid model of the behavior sequence in different time-space domains. Based on the low-level features, the group sparse dictionary is learned based on the low-level features, and the sparse coding of the low-level features is obtained. The spatial average pooling and temporal maximum pooling are used to integrate the coding, so as to obtain high-level features as the descriptor of the final behavior sequence. This kind of descriptor can effectively preserve the spatiotemporal multi-resolution information of human behavior, and at the same time, by eliminating the similar content contained in different behavior categories, a sparse dictionary with stronger expressive power can be obtained to effectively improve the behavior recognition rate.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention belongs to the technical field of computer pattern recognition, and in particular relates to a human behavior recognition method based on depth video sequences. Background technique

[0002] Vision is an important way for human beings to observe and understand the world. With the continuous improvement of computer processing capabilities, we hope that computers can have part of the visual functions of human beings, helping or even replacing human eyes and brains to observe and perceive external things. With the improvement of computer hardware processing ability and the emergence of computer vision technology, people's expectation of computer may become a reality. Human behavior recognition has always been a research hotspot in the fields of pattern recognition, computer vision, and artificial intelligence. The purpose of video-based human behavior recognition is to understand and recognize individual human actions, interactive movement...

Examples

Embodiment Construction

[0011] like figure 1 As shown, this human behavior recognition method based on depth video sequence, the method calculates the four-dimensional normal vector of all pixels in the video sequence, and extracts the pixel points in different layers by constructing the space-time pyramid model of the behavior sequence in different space-time domains. Low-level features, learn group sparse dictionary based on low-level features, obtain sparse coding of low-level features, use spatial average pooling and temporal maximum pooling to integrate coding, and obtain high-level features as the descriptor of the final behavior sequence.

[0012] The present invention constructs a space-time pyramid model, and retains information in the multi-layer space-time field of local descriptors in a targeted manner. At the same time, due to the use of a group sparse dictionary to encode the underlying features, it avoids the interference of different categories containing similar information, making th...