Unlock instant, AI-driven research and patent intelligence for your innovation.

Efficient Face Feature Extraction Method in Unconstrained Environment

A technology of facial features and extraction methods, applied in instruments, computing, character and pattern recognition, etc., can solve problems such as lighting, posture, occlusion and other noise, and achieve reduced feature dimensions, high accuracy, and strong robustness. Effect

Active Publication Date: 2019-10-08
华设设计集团安全科技(江苏)有限公司
View PDF9 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] 1. The traditional global feature extraction algorithm contains redundant information of non-feature regions such as forehead, cheek, and chin, resulting in more noise such as lighting, posture, and occlusion introduced.

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
  • Efficient Face Feature Extraction Method in Unconstrained Environment
  • Efficient Face Feature Extraction Method in Unconstrained Environment
  • Efficient Face Feature Extraction Method in Unconstrained Environment

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0033] The embodiment improves the original HOG operator, constructs 3*3 and 5*5 two-scale main direction rotation HOG gradient operators, and realizes effective face feature extraction in an unconstrained environment. On the one hand, compared with the original HOG gradient operator, the scope of the template of the improved gradient operator is larger, and the number of pixels included is increased, which captures the statistical information of the grayscale change of the face texture from a multi-scale perspective; on the other hand, the improved gradient operator The main direction is rotated counterclockwise every 45° in the range of 0° to 360°, and eight rotated gradient templates are obtained. According to the gradient derivation rule, the eight templates are divided into four groups (Group), and the key point neighborhood ranges are calculated respectively. The gradient orientation histogram in the image describes the statistics of face texture orientation changes from ...

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 present invention provides an effective human face feature extraction method in an unconstrained environment. First, construct a multi-scale main direction rotation gradient template, including two main direction rotation gradient templates of 3*3 and 5*5 scales; The biological vision ROI area in the face image, that is, the region of interest, is marked with key points, and the main direction rotation HOG feature of the 3*3 scale and 5*5 scale is obtained, and the k‑PDR‑HOG‑3 and k‑PDR‑HOG‑5 are combined The features are cascaded and fused to obtain the final multi-scale main direction rotation HOG feature based on key points. Compared with the original HOG algorithm, the multi-scale main direction rotation gradient template constructed by the present invention can extract more abundant and comprehensive face features in an unconstrained environment, and the k-MSPDR-HOG feature has strong robustness and high accuracy .

Description

technical field [0001] The invention relates to an effective face feature extraction method in an unconstrained environment. Background technique [0002] With the development of society and the advancement of technology, human beings have urgent requirements for fast and efficient identity verification technology. Biometric features are the ideal basis for human identity verification, and facial features are currently the most ideal biometric features in identity verification. [0003] At present, due to the good application prospects in attendance, access control system, monitoring system, criminal investigation and other fields, face recognition has attracted more and more researchers' attention. In the past ten years, face recognition technology has made great progress, but the current relatively mature face recognition methods are mostly concentrated in the research of constrained or semi-constrained situations, while in unconstrained environments, face recognition wil...

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 Patents(China)
IPC IPC(8): G06K9/00
CPCG06V40/168
Inventor 童莹严郁黄维曹雪虹
Owner 华设设计集团安全科技(江苏)有限公司