Pose estimation method and device as well as computer system

A technique for estimating position and pose, applied in the field of computer vision, which can solve problems such as learning difficulties, large parameter space, and small effects.

Active Publication Date: 2016-11-16
BEIJING SENSETIME TECH DEV CO LTD
View PDF4 Cites 16 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the parameter space for learning spatial constraints using convolution kernels is too large, making learning very difficult
Also, fo

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
  • Pose estimation method and device as well as computer system
  • Pose estimation method and device as well as computer system
  • Pose estimation method and device as well as computer system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0045] Hereinafter, embodiments of the present application will be described in detail with reference to the detailed description and the accompanying drawings.

[0046] The term "body part" is intended to describe selected parts of the body, preferably parts such as head, neck, shoulders, knees, ankles, such as figure 2 shown. However, it is not intended that the application be limited to the embodiments disclosed herein. For example, more or fewer body parts, or even completely different body parts, may be included for accuracy or efficiency.

[0047] The term "part type" is intended to represent combined information of a body part and adjacent body parts, in particular, the spatial relationship between a body part and adjacent body parts. For example, a wrist connected to an arm in a horizontal state (hereinafter referred to as a horizontal arm) is classified as a wrist of part type 1, and a wrist connected to an arm in a vertical state (hereinafter referred to as a vert...

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 relates to a pose estimation method and device as well as a computer system. The method comprises the following steps of: extracting features corresponding to each body part in a plurality of body parts of an object in an image so as to generate a plurality of feature charts, wherein each body part corresponds to at least one part type; predicting a part type score for each feature in each feature chart, and establishing a plurality of part type score charts according to the predicted part type scores; for at least one point in each part type score chart, optimizing the established part type score charts according to messages related to adjacent body parts of the body part corresponding to the point; and determining an estimated position and an estimated part type of each body part according to the optimized part type score charts, so as to obtain an estimated pose of the object.

Description

technical field [0001] This application relates to computer vision, in particular to a pose estimation method and device, and a computer system. Background technique [0002] Pose estimation of articulated objects is one of the fundamental tasks in the field of computer vision. It solves the problem of part localization of objects in images and has many important applications such as action recognition and human body tracking. The main challenges for this task lie in high articulation, occlusion, clothing, lighting, cluttered background, etc. Recently, deep convolutional neural networks (DCNNs) have been utilized to achieve state-of-the-art performance in the field of human pose estimation. These methods mainly fall into two categories: utilizing DCNN to regress a heat map of each body part location; and learning a deep structured output to further model the relationship between body joints. [0003] DCNN-based heatmap regression models have shown the possibility to learn...

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): G06K9/00
CPCG06V40/25G06V40/103G06T2207/10024G06T2207/30221G06T2207/20084G06T2207/30196G06T2207/20081G06T7/75
Inventor 王晓刚杨巍欧阳万里李鸿升
Owner BEIJING SENSETIME TECH DEV CO LTD
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