Supercharge Your Innovation With Domain-Expert AI Agents!

An action classification method and system under a complex background

A complex background and classification method technology, which is applied in the field of action classification methods and classification systems under complex backgrounds, can solve the problems of low accuracy, environmental changes, and low robustness, and achieve improved accuracy and robustness , the effect of improving accuracy and effectiveness

Active Publication Date: 2019-06-28
NORTHWEST UNIV(CN)
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Solve the problems in the prior art that the single-view image is used for the recognition of human action behavior, which is easily affected by environmental changes, and the method has low robustness and low accuracy.

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
  • An action classification method and system under a complex background
  • An action classification method and system under a complex background
  • An action classification method and system under a complex background

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0047] This embodiment provides a method for classifying actions in complex backgrounds, and the detailed steps are as follows:

[0048] Action information collection: In a closed room, four cameras (EZVIZ CS-C2HC-3B2WFR, viewing angle not lower than 60°) are installed at the four top corners above the human body. An experimenter wears an EEG device (NeuralScanPro+) to move in the middle of the room, and the four cameras installed at the four vertices of the closed room can collect time-series images of the experimenter's human body movements in real time, such as Figure 1-2 shown. Then, a large number of time-series images of human body movements collected are stored, and the time-series images of human body movements are referred to as action time-series images for short in the following. Due to the high resolution of the camera, the size of the captured action time-series images is relatively large. In order to improve the processing speed of the whole method, the size of...

Embodiment 2

[0077] In order to verify the effectiveness of the proposed method, the present invention performs experimental verification on the collected action sequence images:

[0078] The camera collects images under the complex background according to the intensity of three kinds of human body movements. The experiment selects 20,000 images from four angles for classification. The classification experiment follows the standard cross-validation strategy, and uses four-fold cross-validation to carry out the experiment. The experimental results show that the method of the present invention still has a high classification effect under complex backgrounds, with an average accuracy rate of 95%.

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 action classification method and an action classification system under a complex background, and particularly discloses an action classification method and an action classification system based on superpixel segmentation, deep learning and a brain wave signal complex background. moment image features are extracted from the action time sequence image and the action mask image by adopting a CNN class network; the human body action area image characteristics are enhanced; combining the time sequence information and the brain wave form of the image, the human body actionbehavior characteristics are further enhanced, and the problems that in the prior art, a single-view-angle image is used for recognizing the human body action behavior, the human body action behavioris easily affected by environment changes independently depending on the image and under the complex background condition, the recognition rate and accuracy are not high, and robustness is not high are effectively solved.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to an action classification method and a classification system under complex backgrounds. Background technique [0002] Human action recognition is a hot field of artificial intelligence research at present. With the continuous development of the field of intelligent monitoring, audio and video monitoring equipment can monitor all aspects of people's activities and generate a lot of audio and video data that can be used in large quantities. Through these audio and video data, it can be analyzed Human behaviors, judge the intensity of human body movements through the range of human movements, facial expressions and voice tones, and avoid potential criminal activities. [0003] The intensity of human body movements is basically divided into three aspects: violent, calm and soft. The existing research field mainly relies on the collected images to recognize the hu...

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): G06K9/00G06K9/34G06N3/02
CPCY02P90/30
Inventor 曹正文乔念祖卜起荣冯筠
Owner NORTHWEST UNIV(CN)
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

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

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com