首页> 外国专利> FACE IDENTIFICATION METHOD BASED ON AREA DIVISION AND APPARATUS THEREOF

FACE IDENTIFICATION METHOD BASED ON AREA DIVISION AND APPARATUS THEREOF

机译:基于区域划分的人脸识别方法及其装置

摘要

PURPOSE: A face recognizing method based on area division and a device thereof are provided to improve stability and recognition performance by minimizing an influence of face areas excessively getting out of a registered image on identity determination. CONSTITUTION: A feature point extracting unit (10) detects feature points about feature elements characterizing a face shape from an inputted face image. An area dividing unit (20) divides the face image into areas based on the feature points. A feature point comparing unit (30) compares the feature points of the areas with feature points of a registered image in order to classify the areas into a first area and a second area. A feature amount vector obtaining unit (40) obtains a feature amount vector about the first area. A whole feature amount vector obtaining unit (50) deletes areas classified as the second area from the face image and obtains a feature amount vector about the whole face image replaced for a preset molding image. [Reference numerals] (1) Face recognition device; (10) Feature point extracting unit; (20) Area dividing unit; (30) Feature point comparing unit; (40) Feature amount vector obtaining unit; (50) Whole feature amount vector obtaining unit; (60) Similarity calculating unit; (70) Identity determining unit
机译:目的:提供一种基于区域划分的面部识别方法及其设备,以通过最小化过度离开配准图像的面部区域对身份确定的影响来提高稳定性和识别性能。构成:特征点提取单元(10)从输入的面部图像中检测与表征面部形状的特征元素有关的特征点。区域划分单元(20)基于特征点将面部图像划分为区域。特征点比较单元(30)将区域的特征点与注册图像的特征点进行比较,以便将区域分类为第一区域和第二区域。特征量矢量获得单元(40)获得关于第一区域的特征量矢量。整体特征量矢量获得单元(50)从面部图像中删除被分类为第二区域的区域,并获得关于被替换为预设造型图像的整体面部图像的特征量矢量。 [附图标记](1)面部识别装置; (10)特征点提取单元; (20)区域划分单位; (30)特征点比较单元; (40)特征量向量获取单元; (50)整体特征量向量获取单元; (60)相似度计算单元; (70)身份确定单元

著录项

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

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