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

Cross-view-angle action identification method and system based on time sequence information

An action recognition and time sequence technology, applied in the field of pattern recognition, can solve problems such as lack of supervision information, and achieve the effect of solving the problem of apparent information difference

Active Publication Date: 2014-12-10
中科海微(北京)科技有限公司
View PDF2 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Therefore, the present invention proposes a new action video feature based on time-series accumulation of motion intensity. At the same time, the problem of lack of supervision information in the unlabeled mode can be solved by using a weakly supervised cross-view metric learning method. Coarse-grained labeling of the target view video, and then use the coarse-grained labeling information of the source view and the target view to perform cross-view measurement learning, obtain a cross-view measurement method, and apply it to the final classification and recognition

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
  • Cross-view-angle action identification method and system based on time sequence information
  • Cross-view-angle action identification method and system based on time sequence information
  • Cross-view-angle action identification method and system based on time sequence information

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0054] The specific implementation manners of the present invention will be described in detail below in conjunction with the drawings and embodiments.

[0055] The following are general steps of the present invention, specifically as follows:

[0056] First execute step 101, detect the interest point of the motion video, and extract the motion intensity of the interest point; execute step 102, carry out time series accumulation of motion intensity to obtain the motion feature description of the motion video; execute step 103, construct a coarse-grained Classification, coarse-grained classification of the target perspective, so that the original unmarked target perspective video has coarse-grained labeling information; perform step 104, use the coarse-grained labeling information to perform metric learning, and obtain a cross-viewpoint measurement method; perform step 105, use Cross-view metrics for object view classification and weakly supervised cross-view motion recognition...

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 cross-view-angle action identification method and system based on time sequence information and relates to the technical field of mode identification. The method includes detecting interest points of videos and extracting motion intensity of the interest points, wherein the videos include a source view angle video and a target view angle video; conducting time sequence accumulation on the motion intensity according to the time sequence information of the videos to obtain motion characteristic description of the videos; conducting coarseness labeling on the target view angle video according to the motion characteristic description and source coarseness labeling information of the source view angle video to obtain target coarseness labeling information; conducting measurement learning on the source view angle video and the target view angle video through a measurement learning method according to the source coarseness labeling information and the target coarseness labeling information to obtain a cross-view-angle measurement method; conducting action classification on action in the target view angle video through the cross-view-angle measurement method to finish cross-view-angle action recognition.

Description

technical field [0001] The present invention relates to the technology in the field of pattern recognition, in particular to a method and system for cross-viewpoint action recognition based on time series information. Background technique [0002] With the reduction of camera costs and the improvement of people's awareness of public security, video surveillance has received more and more attention. However, a single view has limitations and is constrained by the scene structure, so multi-view video surveillance emerges as the times require. Such as figure 2 As shown, there are video and action category data of a certain view (source view), how to use this information to classify and recognize action videos of another view (target view). Usually, mainstream action recognition frameworks such as image 3 shown. [0003] Cross-view action recognition is an interdisciplinary research field of computer vision, image processing, pattern recognition, machine learning, artificial...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62G06K9/46
Inventor 秦磊刘艺黄庆明成仲炜
Owner 中科海微(北京)科技有限公司
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