基于Log-Gabor和AdaBoost的人脸识别模块设计

Face Recognition Module Design and Experiment Research Based on Log-Gabor and AdaBoost

采用Log-Gabor变换和AdaBoost分类器算法,开发设计具有人脸识别的功能模块.以ORL、YALE和FERET人脸数据库数据以及实地采集图像为实验样本,通过大量直接训练实验和5子集交叉校验,验证了将人脸识别应用于真实环境的可行性和有效性.

In view of widely used characteristics identification requirement, functional modules are designed and developed to recognize human faces based on the Log-Gabor wavelet filters and Adaboost algorithm. By using the face recognition database of ORL,YALE and FERET as well as the experimental samples in the field, the experimental results are obtained through a great number of directly training experiments and cross validation of 5 subsets. In addition, the results are also obtained by analyzing the experimental data of recognition rate curves of Adaboost algorithm. Taking all into consideration, it is feasible and valid to apply the face recognition into the real environment.