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

Human body posture estimation method based on instance segmentation

A technology of human posture and human contour, applied in the field of computer vision, can solve problems such as confusion and dependence on multi-scale features, and achieve the effect of improving accuracy, solving occlusion problems, and good joint positioning

Pending Publication Date: 2022-02-11
赵鸿杰
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

On the one hand, existing methods can cause confusion when people overlap heavily; on the other hand, detection of occluded joints relies more on multi-scale features

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 posture estimation method based on instance segmentation
  • Human body posture estimation method based on instance segmentation
  • Human body posture estimation method based on instance segmentation

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0034] In this embodiment, a human body pose estimation method based on instance segmentation mainly uses the Simplebaseline network and the PRFSM feature selection module to add the human body instance segmentation information obtained after Mask-RCNN processing into the human body pose estimation, and use the human body instance segmentation information Auxiliary positioning to obtain more accurate pixel coordinates of joint points, such as figure 1 As shown, the specific steps are as follows:

[0035] Step 1. Obtain N image datasets with pixel-level labels, denoted as S={S 1 ,S 2 ,...,S n ,...,S N}, where S n Represents the nth image, let the nth image S n H n , and H n ={H n,1 ,H n,2 ,...,H n,t ,...,H n,T},H n,t Indicates the nth image S n label H n in the t-th joint pixel; t∈[1,T], T represents the number of joint pixels; this implementation uses the public image dataset MS COCO for training and testing, which contains a variety of challenging picture scenes...

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 posture estimation method based on instance segmentation. The human body posture estimation method comprises the following steps of: 1, acquiring an image data set with human body joint point labels; 2, performing segmentation processing on the image data set by using a pre-trained Mask-RCNN network to obtain a human body contour information set; 3, constructing a G-line network model, and acquiring joint features through the G-line network model; 4, performing Gaussian transformation on the joint features to obtain a thermodynamic diagram of the joint points; and 5, positioning human body joint points through the thermodynamic diagram of the joint points. According to the method, more precise joint point positions can be obtained, so that the positioning precision is improved, and the defect of large joint point positioning deviation of an existing joint point positioning algorithm is overcome.

Description

technical field [0001] The invention relates to the technical field of computer vision, and specifically designs a human body pose estimation method based on instance segmentation. Background technique [0002] 2D human pose estimation is to identify and localize the joint points of all the people in the image. It is a fundamental and challenging topic in computer vision, and an essential step for computers to understand human actions and behaviors. [0003] In recent years, methods for human pose estimation using deep learning have been proposed one after another, and have achieved performance far exceeding traditional methods. In the actual solution, the estimation of the human body pose is often transformed into the prediction of the joint points of the human body, that is, the position coordinates of each joint point of the human body are first predicted, and then the spatial position relationship between the joint points is determined according to the prior knowledge, ...

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): G06V40/10G06V40/20G06V10/46G06N3/04G06N3/08G06T7/11
CPCG06T7/11G06N3/08G06T2207/20016G06N3/045
Inventor 赵鸿杰
Owner 赵鸿杰
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