首页 | 本学科首页   官方微博 | 高级检索  
     检索      

固定背景下人脸图像的特征提取
引用本文:文可进,蔡学敬,王丽芳.固定背景下人脸图像的特征提取[J].四川职业技术学院学报,2012,22(1):153-156.
作者姓名:文可进  蔡学敬  王丽芳
作者单位:重庆大学城市科技学院电气信息学院,重庆永川,402167
摘    要:对于复杂人脸模式、脸部特征的提取,是人脸自动识别技术的关键,本文对固定背景下人脸图像特征的定位和提取算法进行了分析.文中首先对彩色图像背景进行分割,得到二值化图像;然后采用边界搜索方法确定人脸外接矩形;最后在色度空间中,结合SUSAN角点检测方法,实现人眼、嘴角定位,完成了人脸图像特征的提取和识别.实验证明,该算法在降低运算区域的同时,降低了运算复杂度和其他干扰因素,并且在速度、效率、准确性方面均有良好的性能.

关 键 词:人脸识别  边界搜索  肤色分割  外接矩形  SUSAN角点检测

On FeatureExtractionofFace. Image Under theFixedBackground
WEN Kejin,CAI Xuejing,WANG Lifang.On FeatureExtractionofFace. Image Under theFixedBackground[J].Journal of Sichuan Vocational and Technical College,2012,22(1):153-156.
Authors:WEN Kejin  CAI Xuejing  WANG Lifang
Institution:(Chongqing University,Yongchuan Chongqing 402167)
Abstract:The extraction of complex face pattern and facial feature is the key to automatic face recognition technology.This article analyzes face image feature location and extraction algorithm under the fixed background.In this paper,we finish the face image feature extraction and recognition through the following procedures: we firstly obtain two value images from color image background segmentation.Then we use the boundary search method to determine the external rectangle of face.Finally,we realize human eye and mouth corner location based on SUSAN corner detection method in the color space.We expect to reduce the computational complexity and other interference factors,at the same time,to reduce the operational area.The experimental results also prove that this comprehensive algorithm has good performance on speed,efficiency and accuracy.
Keywords:Recognition of Face  Boundary Search  Skin Color Segmentation  Bounding Rectangle  SUSAN Corner Detection
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号