目 录
摘要2
关键词2
Abstract3
Keyword3
1人脸检测的意义，国内外的情况综述4
2 基于肤色的人脸定位6
2.1彩色空间，以及彩色空间的转换6
2.2人脸肤色模型，二值化7
2.3后处理9
2.4人脸区域的定位10
3视频流10
3.1 视频工具包Video for Windows10
3.2视频流采集的实现11
3.2.1 AVICap主要功能11
3.2.2 视频采集算法12
3.3视频流的显示15
4试验数据与展望17
4.1试验数据17
4.2展望17
参考文献18
附录A 系统使用说明18
摘要
人脸识别技术可以应用于基于网络的身份认证，我们实现了基于WebCam的人脸识别与跟踪系统。本文以WebCam采集的视频流为数据源，截取视频流中的单帧图像，通过转换色彩空间、人脸肤色建模、后处理操作和人脸定位算法实现了人脸检测，并以此为基础实现了在视频流中对于人脸的跟踪。试验结果表明，我们所实现的人脸识别算法适用于近距离人脸的检测，可以应用于基于WebCam的身份认证。
关键词
人脸识别，人脸检测，肤色，定位，视频流
Abstract
Face recognition can be applied to identity authentication on Internet, now we have realized face recognition and tracking system based on WebCam. This essay takes video flowing collected from WebCam as data source, and cuts off Single frame in video flowing. It has realized face detection through transforming colored space, complexion modeling, aftertreatment and face orient algorithms, and based on which the essay has realized face tracking in video flowing.The result of the test shows that the face detection algorithms we realized adapts to short distant face detection, and also can be used as identity authentication based on WebCam.
Keyword
Face recognition, face detection, complexion, orientation, video flowing
