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

Method for recognizing and retrieving action semantics in video

A technology in motion semantics and video, applied in character and pattern recognition, instruments, computer components, etc., can solve problems such as large amount of calculation, and achieve the effect of improving accuracy and reducing calculation amount

Active Publication Date: 2021-11-09
JIANGSU AUSTIN OPTRONICS TECH
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

If we choose to use a time span greater than the maximum value, that is, greater than 60 frames to smooth the calculation and understand all actions, then the amount of calculation will be very large

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
  • Method for recognizing and retrieving action semantics in video
  • Method for recognizing and retrieving action semantics in video
  • Method for recognizing and retrieving action semantics in video

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0048] On the basis of the original SlowFast algorithm, the present invention proposes to determine the input image of the slow channel according to the image stability index and improve the detection accuracy of the slow module; rely on the fast detection of skeletal motion to determine the input video segment of the fast channel and reduce the calculation amount of the fast channel hybrid algorithm.

[0049] A method for action semantic recognition retrieval in a video of the present invention, the video adopts V={Im(f i )}, where Im is the image, f i It is for the image from 1~F imax number of F imax is the maximum number of frames of video V. That is Im(f i ) represents the number f in V i images, such as figure 1 As shown, a method for semantic recognition and retrieval of actions in a video includes the following steps:

[0050] Step 1, using the OpenPose toolbox to extract the key points of the human skeleton in the video image to obtain the three-dimensional coo...

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 method for recognizing and retrieving action semantics in a video, which comprises the following steps of: cutting video into segmented action videos with action, and finding out stable frames in the segmented action videos; carrying out SlowFast self-adaptive cross-frame action recognition; taking the extracted segmented action videos as the input of a Fast algorithm module; and taking the stable frames as the input of a slow algorithm module, performing action semantic recognition by using a SlowFast algorithm to obtain a corresponding action semantic recognition result Out1, establishing a video retrieval library, and when a user inputs a corresponding action semantic query, extracting a corresponding segmented action video for the user to query. According to the method, through preprocessing, the accuracy of the method can be improved while the calculation amount of the SlowFast algorithm is greatly reduced.

Description

technical field [0001] The invention belongs to the technical field of motion semantic recognition, and in particular relates to a method for motion semantic recognition and retrieval in videos. Background technique [0002] In daily life, people sometimes need to find a set of specific action segments in a very long video. For example, in several days of video data, judge the time when the old man fell, so as to observe the surrounding conditions when he fell. However, we probably don't know the specific time and place, or which camera's video this action appeared in. One needs an action-based semantic video retrieval function. When we retrieve the same action in many videos in many places, we can gather these action videos to form an overall effect of the same action, which can be displayed on a multi-screen intelligent display system to play a role. Uniform effect. [0003] In similar work, there are works based on face recognition and narration recognition, but there...

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): G06K9/00
Inventor 翟晓东汝乐凌涛凌婧
Owner JIANGSU AUSTIN OPTRONICS TECH
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