| 基于Adaboost算法的人脸检测系统实现 |
| |
| 引用本文： | 张忠,赵峰,孙炜. 基于Adaboost算法的人脸检测系统实现[J]. 信息技术, 2008, 32(7) |
| |
| 作者姓名： | 张忠 赵峰 孙炜 |
| |
| 作者单位： | 上海交通大学微电子学院,上海,200240 |
| |
| 摘 要： | 在嵌入式平台上用软件实现、软硬件协同、全硬件实现三种实现方法,全面考察了Adaboost算法在嵌入式平台的性能.在Xilinx XUPV2P开发板上,软件实现的检测速度是0.5fps,软硬件协同是1fps,而全硬件实现获得了80fps的检测速度,完全达到实时需求.
|
| 关 键 词： | Adaboost 嵌入式 实时 软硬件协同 FPGA |
Implementation of face detection system based on Adaboost algorithm on embedded system |
| |
| ZHANG Zhong,ZHAO Feng,SUN Wei. Implementation of face detection system based on Adaboost algorithm on embedded system[J]. Information Technology, 2008, 32(7) |
| |
| Authors: | ZHANG Zhong ZHAO Feng SUN Wei |
| |
| Affiliation: | ZHANG Zhong,ZHAO Feng,SUN Wei (Microelectronics School,Shanghai Jiaotong University,Shanghai 200240,China) |
| |
| Abstract: | This paper has studied the performance of Adaboost algorithm on embedded system by comparison of three kinds of method including software,half-hardware and full-hardware.On Xilinx XUPV2P development board,0.5fps is achieved by the implementation of software method.And,the speed of half-hardware design is 1fps.Finally,through full-hardware method,the speed of detection is 80fps which can fulfill the real-time demand. |
| |
| Keywords: | Adaboost embedded system real-time half-hareware co-design FPGA |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
