首页> 外文会议>International Conference on Electronic Design and Signal processing >FACE RECOGNITION USING DCT SIGN ONLY CLASSIFICATION
【24h】

FACE RECOGNITION USING DCT SIGN ONLY CLASSIFICATION

机译:使用DCT标志的人脸识别仅分类

获取原文

摘要

In this paper, a novel face recognition scheme termed as DCT Sign Only Classification (DSOC) is presented. The proposed algorithm uses discrete cosine transform (DCT) sign information of locally normalized frontal face images for feature extraction and a multi-class kernel perceptron (KP) for classification. Experimental results on Yale face database and AR database show that, especially for images with varying illumination the proposed approach gives a significant improvement in identification rate when compared to conventional unconstrained minimum average correlation energy (UMACE) filter and direct linear discrimenant analysis (DLDA) approach.
机译:本文称为DCT符号仅称为分类(DSOC)的新型人脸识别方案。所提出的算法使用离散余弦变换(DCT)用于特征提取的局部归一化正面图像的载体信息,以及用于分类的多级内核Perceptron(KP)。耶鲁脸部数据库和AR数据库的实验结果表明,对于具有不同照明的图像,所提出的方法在与传统的无约束最小平均相关能量(UMACE)滤波器相比,识别率的显着改善,并直接线性歧视分析(DLDA)方法。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号