首页 | 本学科首页   官方微博 | 高级检索  
     

基于DSP的实时人脸检测系统
引用本文:韩相军,关永,王雪立,王万森. 基于DSP的实时人脸检测系统[J]. 微计算机信息, 2005, 0(25)
作者姓名:韩相军  关永  王雪立  王万森
作者单位:北京首都师范大学信息工程学院,北京首都师范大学信息工程学院,北京首都师范大学信息工程学院,北京首都师范大学信息工程学院
基金项目:北京市自然科学基金(4032009)
摘    要:阐述了硬件设计思想、软件开发流程及人脸检测算法的实现。首先以TMS320DM642建立实时图像处理的硬件系统,通过CCD获取视频序列,对当前图像进行降噪、图像增强以及边缘检测的预处理,再利用基于肤色和差分帧相结合的算法对视频序列进行检测、跟踪人脸。本系统所有算法均使用C和汇编语言混合编程并在TIDM642EVM上开发实现。实验结果表明,本系统的算法简单、快速、鲁棒性强。

关 键 词:DSP  人脸检测  人脸跟踪

Real Time Human Face Detection System Based On DSP
Han,Xiangjun Guan,Yong Wang,Xueli Wang,Wansen. Real Time Human Face Detection System Based On DSP[J]. Control & Automation, 2005, 0(25)
Authors:Han  Xiangjun Guan  Yong Wang  Xueli Wang  Wansen
Affiliation:(College of Information Engineering,Capital Nor- mal University,Beijing,100037) Han,Xiangjun Guan,Yong Wang,Xueli Wang,Wansen
Abstract:The designing ideas of the hardware and developing flow of the software and the algorithm of human face detection are expounded. At first setting up the real- time video process- ing system of the hardware using TMS320DM642, obtaining video sequence with CCD, carrying on noise reduction, image strengthen and edge detection that pretreatment measure to cur- rent image, human face detection and track are combined using based- on skin- color and difference frame algorithm. All algo- rithms of the system use C and assembler language mix- program and develop and realize on TI DM642EVM. Experiments show that the algorithm is simple, fast and robust.
Keywords:DSP   Human face detection   Human face track
本文献已被 CNKI 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号