摘要:人脸识别在中国的发展可以追至二十世纪九十年代,到目前为止已经越来越趋于成熟,它覆盖了人工智能,模式识别,数学,数字图像处理等多门学科。该技术可用于身份确认,身份鉴别,访问控制,人机交互,安全监控等等,因其技术特征而使这项技术具有广泛的市场应用前景。本文主要要解决的问题是,从有复杂背景的图像中检测并分割出人脸,再与模板人脸进行相似度检测,确定模板人脸是不是存在与复杂图像中。主要工作分为两个部分:一是人脸的检测和分割,这一部分选用基于统计Adaboost人脸检测算法,利用opencv相关接口实现。二是在分割的人脸区域进行特征的提取,这里采用基于几何特征的算法,定位眼睛,鼻子和嘴角,利用这些位置构造特征向量,利用这些向量完成匹配。65261
毕业论文关键词:人脸检测 Adaboost 灰度投影 几何特征
Title Template-based complex adaptive image segmentation and Matching
Abstract
Face recognition in China's development can chase the 1990s,So far has become more and more mature.It covers pattern recognition, artificial intelligence, mathematics, digital image processing and many other subjects.The technology can be used for identity recognition, authentication, access control, security monitoring, human-computer interaction, etc..Leaving its technical features of this technology has a broad market prospect。In this paper, the problem is to be solved,The image from a complex background to detect and segment the face,Then the similarity with the template of face detection,Sure there is not a template face image with complex.Major work is pided into two parts: First, face detection and segmentation, this part of the selection based on statistical Adaboost face detection algorithm, using OpenCV related interface..Second, in the face region segmentation for feature extraction, where the use of algorithms based on geometric features, positioning the eyes, nose and mouth, the use of these locations feature vectors constructed using these vectors complete match.
Key words Face Detection Adaboost Gray Projection Geometric characteristics
目次
1 绪论 1
1.1人脸识别研究概述 1
1.2人脸识别的应用领域 1
1.3人脸识别技术研究现状 1
1.3.1 国外研究现状 1
1.3.2 国内研究现状 2
1.4 人脸识别的构成 2
1.4.1图像的获取 3
1.4.2人脸的检测与定位 3
1.4.3图像预处理 3
1.4.4特征的选择和提取 4
1.4.5识别 4
1.5 人脸识别的理论及方法研究 4
1.5.1基于主成分分析的人脸识别——特征脸法(PCA)。 4
1.5.2基于BP神经网络 5
1.5.3 基于弹性模型 6
1.5.4 其他方法 6
1.6 本论文的目的和内容 6
2 人脸检测 Adaboost模板的复杂图像的自适应分割与匹配研究:http://www.youerw.com/jisuanji/lunwen_72807.html