Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Video behavior recognition method and system based on hierarchical dynamic depth projection difference image representation

A difference image and recognition method technology, applied in character and pattern recognition, instruments, biological neural network models, etc., can solve the problem of ignoring timing information, and achieve the effect of improving performance and excellent recognition results

Active Publication Date: 2019-03-12
SHANDONG UNIV
View PDF5 Cites 18 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although the DMM includes the motion change information of the behavior video, the sum operation ignores the timing information of the behavior in the video

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Video behavior recognition method and system based on hierarchical dynamic depth projection difference image representation
  • Video behavior recognition method and system based on hierarchical dynamic depth projection difference image representation
  • Video behavior recognition method and system based on hierarchical dynamic depth projection difference image representation

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0065] In one or more implementations, a video behavior recognition method based on hierarchical dynamic depth projection difference image representation is disclosed, such as figure 1 As shown, the spatiotemporal dynamic information of behaviors in videos can be extracted simultaneously from different time scales.

[0066] For a depth video sequence, firstly, the corresponding depth projection image sequence is obtained by projection in three orthogonal Cartesian planes, and the depth projection image sequence in each plane is sampled hierarchically in time order to generate different time scales. A sequence of sampled depth projection maps for .

[0067] The depth projected difference image (Depth Projected Difference Image, DPDI) sequence is obtained by calculating the absolute frame difference between two adjacent frames in the depth projection image sequence at each time scale.

[0068] The DPDI sequences under different time scales in the three projection planes can com...

Embodiment 2

[0151] A video behavior recognition system based on layered dynamic depth projection difference image representation disclosed in one or more implementations includes a server, the server includes a memory, a processor, and is stored on the memory and can run on the processor A computer program, the processor implements the video behavior recognition method based on layered dynamic depth projection difference image representation described in Embodiment 1 when executing the program.

Embodiment 3

[0153] A computer-readable storage medium disclosed in one or more implementation manners, on which a computer program is stored, and when the program is executed by a processor, the video representation based on layered dynamic depth projection difference image described in Embodiment 1 is executed. Behavior recognition method.

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a video behavior recognition method and system based on hierarchical dynamic depth projection difference image (HDDPDI) representation, Firstly, the depth video sequences are projected into three orthogonal Cartesian planes to generate corresponding depth projection image sequences to capture 3D contour and motion characteristics of human behavior, and the corresponding HDDPDI is constructed in each projection plane based on the depth projection image sequences. HDDPDI can simultaneously encode the spatio-temporal motion dynamics of behavior in video. CNN can automatically learn the difference features in the image. In order to verify the effectiveness of the proposed HDDPDI video representation, a CNN-based behavior recognition framework is constructed, in which three behavior classification schemes are designed. The HDDPDI in the three projection planes are separately inputted into three identical pre-trained CNNs to fine-tune the network parameters, Differentclassification schemes use different network layers of CNN to compare their effects on behavior recognition. Each classification scheme combines the information of three projection planes to obtain aricher and more comprehensive representation of behavior features.

Description

technical field [0001] The present invention relates to the technical field of behavior recognition, in particular to a video behavior recognition method and system based on hierarchical dynamic depth projection difference image representation. Background technique [0002] In recent years, human action recognition has attracted more and more attention in the field of computer vision. Traditional behavior recognition methods based on RGB data usually focus on human body contour features, key poses, etc. Although they may have high recognition performance in some specific application backgrounds, RGB-based behavior recognition methods are very sensitive to changes in lighting conditions, and are not suitable for more challenging scenes (with occlusions and cluttered backgrounds). ) human behavior cannot be accurately identified. [0003] The emergence of low-cost integrated depth sensors such as Microsoft's KinectTM can simultaneously capture RGB (red, green, blue) video an...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06K9/00G06K9/62G06N3/04
CPCG06V20/42G06V20/46G06N3/045G06F18/2411
Inventor 马昕武寒波荣学文宋锐田新诚田国会李贻斌
Owner SHANDONG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com