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

Behavior image classification method based on human body local semantic knowledge

A classification method and technology of human body parts, applied in the fields of image recognition and artificial intelligence, can solve problems such as performance bottlenecks and large modal differences, and achieve the effect of improving accuracy and behavior recognition.

Active Publication Date: 2021-09-28
SHANGHAI JIAO TONG UNIV
View PDF9 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Existing image behavior detection methods directly infer human behavior from image-level features. Due to the large modal difference between the two, this type of method is prone to performance bottlenecks

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
  • Behavior image classification method based on human body local semantic knowledge
  • Behavior image classification method based on human body local semantic knowledge
  • Behavior image classification method based on human body local semantic knowledge

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0021] Such as figure 1 As shown, the present embodiment relates to an object-attribute combination image recognition method based on symmetry and group theory, comprising the following steps:

[0022] Step 1, build a dataset: use a public image dataset containing human behavior, and obtain the human bounding box b h , object bounding box b o (When the behavior is a human-object interaction behavior) and the behavior label label action , define the human body part behavior state of the people participating in the interaction, and finally get 76 different human body part states; based on these definitions, mark the human body part state of each human behavior instance in the image data set, and get the following results: Human body part state label pasta And the human body part attention vector label with a length of 10 att , to represent whether each part contributes to the behavior sample; and perform two-dimensional human body pose estimation on the people in the trainin...

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 an image classification method based on human body local behavior semantic knowledge. The method comprises the following steps: establishing a human body part behavior state recognition model for obtaining human body local fine-grained semantic representation, and performing model training; then converting visual information in an image to be detected into language-based priori knowledge by utilizing natural language understanding, fusing the priori knowledge and the visual information to generate a fine-grained behavior representation vector, and migrating the fine-grained behavior representation vector to a computer visual behavior and recognition task; and finally, reasoning the overall behavior by combining the local fine-grained features of the human body to complete a behavior understanding process to obtain a classification result. According to the method, very ideal recognition performance improvement is achieved in various complex behavior understanding tasks; and meanwhile, the method has the advantages of one-time pre-training and multiple-time diverse migration, and has generalization and flexibility.

Description

technical field [0001] The invention relates to a technology in the field of image recognition and artificial intelligence, in particular to an image classification method based on semantic knowledge of human body partial behavior. Background technique [0002] Human behavior detection is an important branch of computer vision, and its goal is to infer human behavior and interaction with the environment in images or videos. Behavior detection is widely used in the fields of intelligent driving, security, and robotics. It is one of the most important artificial intelligence technologies for the industry and has attracted more and more attention. Machine learning mainly studies computer algorithms that can be automatically improved through experience. It usually obtains, abstracts, and summarizes key information and knowledge from a large amount of empirical data. Artificial neural networks are an important branch of machine learning and are currently widely used in 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
IPC IPC(8): G06K9/00G06K9/32G06K9/62G06N3/08
CPCG06N3/084G06F18/253
Inventor 李永露徐良刘欣鹏许越卢策吾
Owner SHANGHAI JIAO TONG 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