人脸识别 PCA 算法的硬件实现

摘要

In order to satisfy the real-time capability of face recognition system,a new method is proposed to realize the face rec-ognition system based on FPGA hardware platform.The principal component analysis (PCA)algorithm is the theoretical basis in this system and Xilinx Company’s System Generator software is the development tool.After debugging,simulating and analyzing the computation load and sequential logic,the result shows that the new method can reduce the calculating time compared with MATLAB simulation.%为满足人脸识别系统的实时性，提出了一种基于 FPGA 硬件平台的人脸识别系统的实现方法。以传统的主成份分析算法为理论基础，利用 Xilinx 公司的 System Generator 开发环境实现算法的硬件逻辑。通过调试和仿真后，分析运算负荷和时序逻辑。结果表明，使用 Spartan-3A DSP XC3SD3400A FPGA 开发套件，在保证识别率的前提下，硬件实现的算法处理速度要远快于 MATLAB 中的算法实现。