Human body behavior identification method and system based on multi-target detection 3D CNN

A recognition method and multi-target technology, applied in the field of human behavior recognition method and system based on multi-target detection 3DCNN, can solve the problems of increased labor costs, inability to transfer, poor model generalization ability, etc., achieve the ability of reinforcement learning, reduce effect of difficulty

Active Publication Date: 2019-07-05
SOUTH CHINA UNIV OF TECH
View PDF6 Cites 22 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The design of artificial features is often only carried out for a specific part of the data, which leads to poor generalization ability of the model and cannot be quickly migrated to other applications, which greatly increases the cost of labor

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
  • Human body behavior identification method and system based on multi-target detection 3D CNN
  • Human body behavior identification method and system based on multi-target detection 3D CNN
  • Human body behavior identification method and system based on multi-target detection 3D CNN

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0052] The present invention will be further described below in conjunction with specific examples.

[0053] see figure 1 As shown, the human behavior recognition method based on multi-target detection 3D CNN provided in this embodiment includes the following steps:

[0054] 1) Establish a human behavior recognition data collection system to obtain human behavior video datasets. Here, the public datasets are mainly used for model training, and the test datasets are collected by cameras in real environments;

[0055] 2) Convert the collected video datasets into frame datasets and use the SSD (full name: Single ShotMultiBox Detector) detection algorithm to calibrate and crop datasets;

[0056] 3) Establish a 3D CNN learning model, learn the data sets separately, and fuse the learned features;

[0057] 4) Utilize the Softmax classifier to classify and identify the features after fusion;

[0058] 5) Classification and calibration identification or early warning report for class...

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 human body behavior identification method and system based on multi-target detection 3D CNN, and the method comprises the steps: 1) carrying out the preprocessing of a video,and enabling the video flow to be converted into an image frame; 2) carrying out calibration cutting on a target object in the video by adopting a relatively mature SSD detection technology at present; 3) establishing a feature extraction network structure of the image frame data and the calibration clipping data; 4) establishing a feature fusion model, and fusing the two features extracted in the step 3); 5) carrying out classification by using a Softmax regression model classifier; and 6) carrying out fine tuning on the trained model according to an actual application scene or a public dataset. According to the method, a condition of information loss caused by convolution of the current deep neural network model in the time dimension is made up, the expression of features in the time dimension is enhanced, the recognition efficiency of the model is integrally improved, and the model can better understand the behavior action of a human body.

Description

technical field [0001] The present invention relates to the technical field of human behavior recognition and analysis, in particular to a human behavior recognition method and system based on multi-target detection 3D CNN. Background technique [0002] Human behavior recognition refers to the recognition of human behavior or actions in the real environment, which can be applied in various fields. At present, common application scenarios include: intelligent monitoring, smart home, human-computer interaction and human behavior attribute analysis, prediction and other fields. However, improving the accuracy and efficiency of recognition is still a very challenging task and has attracted extensive attention from all researchers. [0003] In the past few decades, the extraction and representation of human behavior features mainly stayed at the manual stage, and the design and extraction of features by humans often depended on the designer's experience. Common artificial featu...

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): G06K9/00G06N3/04G06N3/08
CPCG06N3/08G06V40/20G06V20/41G06V20/46G06V2201/07G06N3/045
Inventor 董敏李永发毕盛聂宏蓄
Owner SOUTH CHINA UNIV OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap