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

Method for recovering occluded human body posture sequence based on deep learning

A technology of human posture and deep learning, applied in neural learning methods, instruments, biological neural network models, etc., can solve problems such as inaccurate posture and achieve good visual effects

Pending Publication Date: 2020-10-02
SOUTH CHINA UNIV OF TECH
View PDF6 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The method at this stage is to use the prediction method: to predict the unknown target pose sequence through the pre-order pose sequence, but only considering the information of the pre-order pose may lead to the prediction of the generated pose is not accurate enough

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 recovering occluded human body posture sequence based on deep learning
  • Method for recovering occluded human body posture sequence based on deep learning
  • Method for recovering occluded human body posture sequence based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0041] A method based on deep learning to restore occluded human pose sequences, such as figure 1 shown, including the following steps:

[0042] S1. Process according to the video data, obtain a human walking data set, perform data preprocessing and division on the human walking data set, and obtain a pre-order posture sequence, a target posture sequence, and a post-order posture sequence;

[0043] S2. Training the first prediction model based on RNN, the model takes the pre-order pose sequence as input, and the occluded pose sequence as output;

[0044] S3, train the second predictive model based on CNN, this model uses post-sequence pose as input, is occluded pose sequence as output;

[0045] S4. Add the outputs of the first prediction model and the second prediction model with weights, and add a MergeNet network to construct a total training model, and obtain a final pose sequence by training the total model.

[0046] The specific process is as follows:

[0047] (S11) Af...

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 recovering an occluded human body posture sequence based on deep learning, and the method comprises the following steps: carrying out the processing according to video data, obtaining a human body walking data set, carrying out the data preprocessing and dividing of the human body walking data set, and obtaining a preorder posture sequence, a target posture sequence and a postorder posture sequence; training a first prediction model based on RNN, the model taking a preorder attitude sequence as an input and an occluded attitude sequence as an output; training a second prediction model based on the CNN, wherein the model uses a subsequent posture as an input and a shielded posture sequence as an output; performing weighted addition on the output of the first prediction model and the output of the second prediction model, adding a MergeNet network, constructing a training total model, and obtaining a final attitude sequence through the training totalmodel; according to the invention, the motion posture sequence of the human body shielded by the obstacle can be recovered by analyzing the information of the preorder posture sequence and the subsequent posture sequence.

Description

technical field [0001] The present invention relates to the research field of multimedia information processing, in particular to a method for restoring occluded human posture sequences based on deep learning. Background technique [0002] Attitude generation is a technology that infers the attitude sequence in the next period of time based on the known attitude sequence information; the method based on deep learning can learn through a large amount of data, so that the model can acquire the ability to infer subsequent attitudes, which is more universal scope of application. The method at this stage is to use the prediction method: to predict the unknown target pose sequence through the pre-order pose sequence, but only considering the information of the pre-order pose may lead to inaccurate predicted poses. Contents of the invention [0003] The main purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art, and to provide a metho...

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): G06K9/00G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06V40/20G06V20/48G06N3/045G06F18/214
Inventor 聂勇伟杨韫韬李桂清
Owner SOUTH CHINA UNIV OF 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