人脸识别系统中人脸检测模块的的研究与实现.doc

人脸识别系统中人脸检测模块的研究与实现 内容摘要：人脸检测是计算机视觉与模式识别领域非常活跃的研究课题，在身份验证、人机界面、可视通信、虚拟现实、公安档案管理、基于内容的图像检索等很多方面都有着广泛的应用。 作为人脸识别的重要的第一步，人脸检测所做的工作是将人脸从图像背景中检测出来，它是人脸识别、视点跟踪和人脸图像压缩等应用中的重要环节。由于受图像背景、亮度变化及人的头部姿势等因素的影响，使得人脸检测成为一项复杂的、具有挑战性的研究课题。 本文论述了人脸检测技术的基本概念，分析和探讨了多种人脸检测方法的基本理论，对人脸检测方法进行了深入的研究和讨论，实验表明本文研究的人脸检测方法是合理的，具有一定的理论价值与实用价值。本文的研究工作主要包括: 基于AdaBoost学习算法，从一个较大的特征集中选择少量关键的haar-like特征，产生一个高效的强分类器。再用级联方式将单个的强分类器级联成为一个更加复杂的级联分类器。针对AdaBoost算法在训练过程中出现的退化问题及样本权重扭曲的现象，本文对样本权重的更新规则作出了适当的调整，在一定程度上避免了退化现象，提高了分类器的性能。实验结果表明，新的人脸检测器检测效果显著加强。 关键词：人脸检测AdaBoost算法 Haar特征  Research and implementation of face detection module in face recognition system Abstract:Human face recognition and detection are the most active and chaallenging tasks for computer vision and pattern recognition.It can be widely applied to such fields as personal identification,human-computer interface,visual communication,criminal archive administration,content-based image retrieval,etc. As the first step of face recognition,the task of face detection is to detect human faces from background of image.However,face detection result is usually affected by the background,brightness or head posture of image and so on,which makes the process of detection more complicated. A great amount of literatures,surveys and research papers concerning up-to-date techniques of face detection and face recognition are and analyzed.Some hot issues about face detection are discussed in this paper.The experiments indicate that the methods of face detection proposed in this paper are reasonable,showing a certain degree of theoretical and practical value.The research work of this paper mainly about: Using face detection method based on AdaBoost learning algorithm,which selects few key haar-like features from a large set of features,to build a robust cascade classifier.Focusing on the disadvantages of classical AdaBoost algorthm,this paper analyses the issues of overfitting and distortion of sample weights in training process and come up with a new me