论文部分内容阅读
随着安防、人机交互等需求的提出，人脸检测已经成为模式识别领域的一个研究热点，日益受到人们的重视。近年来，随着嵌入式系统的快速发展，出现了基于嵌入式设备的人脸检测系统，由于其具有体积小、成本低、布设方便的优点，逐渐得到了人们的青睐，所以取得了很大的发展。本文主要研究和开发了基于德州仪器DM365视频监控设备的人脸检测系统，针对 Adaboost 的人脸检测算法进行了研究，利用改进的带加权判决函数的级联分类器来提高检测准确率，并给出了在DM365嵌入式视频监控设备上的实现。“,”The face detection algorithm, currently, receives an increasing attention, and has become a hot topic in the field of the pattern recognition with the requirement of the security guards against theft and the human-computer interaction. With the rapid development of the embedded system, the face detection system based on embedded platform emerged in recent years. And this system has earned the favor of users and developed rapidly due to its advantages of tiny volume, low cost and convenient layout. In this paper, we researched and developed a face detection system based on DM365. The AdaBoost algorithm was improved by employing a cascade classifier with weighted decision function. This improved algorithm enhanced accuracy rate of the detection and was implemented on DM365.