本科毕业设计论文-基于移动窗口的人脸图像识别方法研究.doc

河北大学2009届本科生毕业论文(设计) PAGE 基于移动窗口的人脸图像识别方法研究 摘 要 人脸识别是机器视觉和模式识别领域最富有挑战性的课题之一，同时也具有广泛的应用意义。目前在基于正面静态人脸检测与识别、人脸特征的提取等方面已取得了一些可喜的成果。但是,在较复杂的环境下，诸如对人脸表情的识别、光照补偿与光照模型的建立、多种检测数据的融合等方面还缺乏着实有效的方法。 本文介绍了移动窗口，移动窗口广泛应用于图像处理、模式识别和数字信号处理中，它具有数据量大、计算密集等特点。移动窗口针对图像的算法的一般描述是：在规模为 W×H 的图像中，按一定规律移动 w×h 的窗口(W>>w， H>>h)，对窗口内像素点的像素值进行一系列运算，运算结束后窗口向右或向下移动一步，直到完成对整幅图像的处理。用移动窗口可以使图像边界的细节保持好，处理速度快，从而提高了人脸图像识别的准确率。 关键词：人脸识别 移动窗口 人脸特征检测 ABSTRACT Face recognition is the field of machine vision and pattern recognition,one of the most challenging issues,but also the meaning of a wide range of applications. Based on the positive static in face detection and recognition,,facial feature extraction, such as there has been some encouraging results. However, in more complex circumstances, such as facial expression recognition, compensation and light illumination model, the integration of a variety of test data, etc. is also a lack of really effective method. In this paper, moving window is widely used in image processing, pattern recognition and digital signal processing, it has a large amount of data, computing-intensive and so on. Moving window for the image of the general description of the algorithm is: in terms of size W × H for the image, according to a certain law of w × h moving window (W>> w, H>> h),on the window of the pixel value of pixels a series of operations, operations after the end of the window or down a step to the right until the completion of the whole image processing. Moving window can be used to make the details of the image border to maintain a good,fast processing speed, resulting in improved facial image recognition accuracy. Key words: Face recognition MovingWindow Facial feature detection 目 录 TOC \o "1-3" \h \z \u HYPERLINK \l "_Toc231177059" 一 引言 PAGEREF _Toc231177059 \h 1 HYPERLINK \l "_Toc231177060" 1.1人脸图像识别的研究背景及意义 PAGEREF _Toc231177060 \h 1 HYPERLINK \l "_Toc231177061" 1.2人脸图像识别研究的状况和进展 PAGEREF _Toc231177061 \h 1 HYPERLINK \l "_Toc231177062" 1