首页> 外文会议>NAFOSTED Conference on Information and Computer Science >Gender classification by LPQ features from intensity and Monogenic images
【24h】

Gender classification by LPQ features from intensity and Monogenic images

机译:通过强度和单一图像的LPQ功能进行性别分类

获取原文

摘要

Gender is one of the most useful facial attributes which are detected from human face images. In this work, we introduce a new gender classification system based on features extracted by Local Phase Quantization (LPQ) operators from intensity and Monogenic images. More detailed, the LPQ features are obtained from the input image (the intensity one) and from three other Monogenic components in the feature extraction stage. In the classification stage, we employ the binary SVM classifier to predict the gender of the given test images. The comparisons among our experimental results upon two public databases, LFW and Groups dataset, and those of other systems show that the proposed system is comparable with state-of-the-art approaches when it attains competing accuracy rates (97.0% on LFW and 91.58% upon Groups dataset).
机译:性别是从人脸图像中检测到的最有用的面部属性之一。在这项工作中,我们介绍了一种基于局部相位量化(LPQ)运营商从强度和单一图像提取的功能的新的性别分类系统。更详细地,LPQ特征是从输入图像(强度一)获得的,并且在特征提取阶段中的三个其他单一组分中获得。在分类阶段,我们使用二进制SVM分类器来预测给定测试图像的性别。我们在两个公共数据库,LFW和组数据集的实验结果中的比较以及其他系统的比较表明,当竞争精度率(LFW和LFW上的97.0 %)时,所提出的系统与最先进的方法相当小组数据集的91.58 %)。

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