摘要
该文研究了人脸图像频率域的鉴别信息，提出了一种Gabor小波变换与主成分分析相结合的人脸图像特征提取方法。不仅利用了图像中的频域信息，而且利用主成分分析法使特征向量维数有效降低。最后，采用随机森林作为分类器。该文实验结果表明．该方法在小规模及大规模数据集上均具有较好的性能。
This paper considers the discriminantive information in frequency space of face images. A new feature extraction method is proposed based on Gabor Wavelet Transform and Principle Component Analysis. Not only the frequency information is used, and also the di- mensionality is reduced by PCA. Finally, Random Forest is adopted as the classifier. The experiments in the paper valuate the performance of the proposed method on large databases and small databases.
作者
胡锋
邢洁清
HU Feng, XING Jie-qing （Department of Modern Education Technology, Qiongtai Teachers College, Haikou 571100, China）
出处
《电脑知识与技术》
2011年第6期3899-3900,共2页
Computer Knowledge and Technology
基金
海南省教育厅高校科研项目（hjk2010-50）
