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

Video behavior timeline detection method

A detection method and time axis technology, applied in the field of video analysis, can solve problems such as insufficient use of candidate frames and efficiency to be improved

Active Publication Date: 2018-11-16
PEKING UNIV SHENZHEN GRADUATE SCHOOL
View PDF5 Cites 20 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the SSN model itself has two shortcomings: one is that SSN tries to accurately locate the time boundary of the behavior, but ignores the information of the time boundary; The candidate frame is discarded directly, the candidate frame is not fully utilized, and the efficiency needs to be improved

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 timeline detection method
  • Video behavior timeline detection method
  • Video behavior timeline detection method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0029] Below in conjunction with accompanying drawing, further describe the present invention through embodiment, but do not limit the scope of the present invention in any way.

[0030] Video behavior timeline detection is a subproblem of video behavior detection, which mainly focuses on locating the beginning and end of time when behaviors occur within a video. It plays a key technical support role in the fields of intelligent monitoring system, human-computer interaction and video search. Since the video contains a lot of information, most of which are useless noise, it is very important to automatically locate the video segments of human interest on the time axis for subsequent processing before processing. In practical applications, quite high requirements are placed on the accuracy of video behavior positioning. In order to accurately locate the start time and end time of a video action, it is very important to extract the time boundary information of the action. This ...

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 timeline detection method. The method comprises the following steps: performing modeling based on a deep learning and time structure, detecting a video behavior timeline in combination with coarse granularity detection and fine granularity detection, and extracting temporal and spatial characteristics of a video by using a dual-flow model on the basis of anexisting model SSN; modeling a time structure of a behavior, and dividing a single behavior into three stages; then providing a new characteristic pyramid capable of effectively extracting time boundary information of a video behavior; and finally, combining the coarse granularity detection and the fine granularity detection to make a detection result more precise. The video behavior timeline detection method is high in detection precision, and the detection precision is higher than the detection precision of all of the current existing disclosed methods; the video behavior timeline detectionmethod is wide in application range, is applicable to detection of video clips in which people are interested in an intelligent monitoring system or a human-machine supervision system, is favorable for subsequent analysis and processing, and has an important application value.

Description

technical field [0001] The present invention relates to the technical field of video analysis, in particular to a video behavior timeline detection method, which is based on deep learning and combined with video context information to detect the timeline of human behavior in a video. Background technique [0002] Videos containing human behavior can be divided into two categories: one is artificially cropped video containing only human behavior without any extraneous background video; the other is uncropped video after shooting, in which not only Includes human behavior and contains extraneous background segments such as credits, audience, etc. Video behavior timeline detection refers to locating the start time and end time of human behavior in a video that has not been manually cropped, and identifying the category of human behavior. The existing video behavior timeline detection methods mainly follow a two-step strategy: first, extract a large number of video behavior tim...

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): G06K9/00G06K9/46H04N7/18G06N3/04
CPCH04N7/18G06V40/20G06V20/40G06V10/40G06N3/045
Inventor 李革张涛李楠楠林凯孔伟杰李宏
Owner PEKING UNIV SHENZHEN GRADUATE SCHOOL
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