Improved cascaded regression face feature point positioning algorithm

A cascade regression, face feature technology, applied in computing, computer parts, image data processing and other directions, can solve problems such as face alignment failure

Inactive Publication Date: 2021-03-12
CHONGQING UNIV OF POSTS & TELECOMM
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] Aiming at the problem that face alignment fails due to large occlusion and postu...

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
  • Improved cascaded regression face feature point positioning algorithm
  • Improved cascaded regression face feature point positioning algorithm
  • Improved cascaded regression face feature point positioning algorithm

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0015] specific implementation plan

[0016] Regression-based face alignment algorithms predict the coordinates of faces by iteratively updating the initial shape by learning a regression function that directly maps image appearance to target output. The quality of the initial shape has a significant impact on the final alignment accuracy. The algorithm typically uses an average human face as an initial shape, or randomly picks a shape from training images during testing. Due to large occlusions and pose variations, poor initial shapes may lead to local optima for face alignment. And these sparse point sets (5-point coordinates) are the most prominent points on the face, which can achieve good performance even in changes in occlusion, pose and expression, so these 5 feature points can approximate the face shape and pose .

[0017] The system implementation method of the present invention is described in detail below in conjunction with accompanying drawings and specific exa...

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 improved cascaded regression face feature point positioning algorithm, and belongs to the technical field of computer vision. According to the algorithm, initialization andregression of a traditional cascade regression technology are effectively improved, the initial shape of projection is generated by using a field (5-point coordinates), and fine adjustment and correction are carried out, so that the performance based on the cascade regression algorithm is greatly improved; different models are established under different human face pose visual angles, a single model is replaced by multiple models to improve the human face feature point positioning precision, and the shape variance in regression model construction is greatly reduced, so that the learned regression model has better robustness for shape change. Compared with a traditional cascade regression algorithm, the improved cascade regression face feature point positioning algorithm is higher in robustness and has higher practical application values in the technical field of computer vision.

Description

technical field [0001] The present invention aims to design an improved cascade regression face feature point positioning algorithm, which overcomes the problem of face alignment failure due to large occlusion and posture changes by improving the traditional cascade regression algorithm, thereby Significantly improved the performance of the cascade-based regression algorithm. [0002] technical background [0003] Face alignment is one of the hotspots in the field of computer vision. Face alignment refers to the localization of facial landmarks such as eyebrows, eyes, nose tip, and mouth corners. Efficient face alignment is often regarded as a preprocessing step for many vision tasks, such as face recognition, expression analysis, facial makeup, and 3D face modeling, etc. With the rapid development and popularization of multimedia technology, efficient face alignment is very important in multimedia applications. [0004] In recent years, face feature point location has dev...

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): G06K9/00G06T7/41
CPCG06T7/41G06V40/161G06V40/165G06V40/171
Inventor 代少升黄涛
Owner CHONGQING UNIV OF POSTS & TELECOMM
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