首页> 外国专利> AUXILIARY IDENTIFICATION METHOD AND APPARATUS FOR FACE-TO-FACE INTERVIEW DECEPTION, AND ELECTRONIC DEVICE AND STORAGE MEDIUM

AUXILIARY IDENTIFICATION METHOD AND APPARATUS FOR FACE-TO-FACE INTERVIEW DECEPTION, AND ELECTRONIC DEVICE AND STORAGE MEDIUM

机译:用于面对面采访欺骗的辅助识别方法和装置,以及电子设备和存储介质

摘要

An auxiliary identification method and apparatus for face-to-face interview deception, and an electronic device and a storage medium. The auxiliary identification method for face-to-face interview deception comprises: acquiring a face-to-face interview video of a target object, and taking out, from the face-to-face interview video, n target video segments of the target object answering face-to-face interview questions (S31); performing facial action unit intensity extraction on the target object in the n target video segments, so as to obtain a facial action unit intensity sequence of the target object (S32); acquiring a statistical quantity of the facial action unit intensity sequence (S33); and inputting the statistical quantity of the facial action unit intensity sequence into a pre-trained deception identification model, in which two layers of long short-term memory networks are stacked, to carry out processing in a cyclical manner, so as to obtain a deception score of the target object (S34). The method is beneficial to the improvement of the accuracy of auxiliary identification of face-to-face interview deception.
机译:面对面采访欺骗的辅助识别方法和装置,以及电子设备和存储介质。面对面面对面欺骗的辅助识别方法包括:从面对面的面对面视频,从面对面的面对面视频中获取面对面的面对面的面试视频,并取出目标对象的视频片段回答面对面的面试问题(S31);对N个目标视频段的目标对象执行面部动作单位强度提取,以便获得目标对象的面部动作单位强度序列(S32);获取面部动作单位强度序列的统计量(S33);并将面部动作单位强度序列的统计量输入到预先训练的欺骗识别模型中,其中堆叠了两层长的短期存储网络,以便以周期性方式进行处理,以便获得欺骗目标对象的得分(S34)。该方法有利于提高面对面采访欺骗的辅助识别准确性。

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