Video-Based Deception Detection with Non-Contact Heart Rate Monitoring and Multi-Modal Feature Selection

作     者：Yanfeng Li Jincheng Bian Yiqun Gao Rencheng Song Yanfeng Li;Jincheng Bian;Yiqun Gao;Rencheng Song

作者机构：Department of Biomedical EngineeringHefei University of TechnologyHefei 230009China 

出 版 物：《Journal of Beijing Institute of Technology》 (北京理工大学学报（英文版）)

年 卷 期：2024年第33卷第3期

页      面：175-185页

核心收录：

学科分类：030604[法学-侦查学] 03[法学] 08[工学] 0710[理学-生物学] 0831[工学-生物医学工程（可授工学、理学、医学学位）] 1007[医学-药学(可授医学、理学学位)] 0810[工学-信息与通信工程] 1205[管理学-图书情报与档案管理] 100706[医学-药理学] 1002[医学-临床医学] 1001[医学-基础医学(可授医学、理学学位)] 080203[工学-机械设计及理论] 0838[工学-公安技术] 0835[工学-软件工程] 0802[工学-机械工程] 0836[工学-生物工程] 0701[理学-数学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 0306[法学-公安学] 030609[法学-涉外警务学] 

基　　金：National Natural Science Foundation of China(No.62271186) Anhui Key Project of Research and Development Plan(No.202104d07020005) 

主　　题：deception detection apparent visual features remote photoplethysmography non-contact heart rate feature selection 

摘      要：Deception detection plays a crucial role in criminal *** contain a wealth of information regarding apparent and physiological changes in individuals,and thus can serve as an effective means of deception *** this paper,we investigate video-based deception detection considering both apparent visual features such as eye gaze,head pose and facial action unit(AU),and non-contact heart rate detected by remote photoplethysmography(rPPG)*** wrapper-based feature selection methods combined with the K-nearest neighbor(KNN)and support vector machine(SVM)classifiers are employed to screen the most effective features for deception *** evaluate the performance of the proposed method on both a self-collected physiological-assisted visual deception detection(PV3D)dataset and a public bag-oflies(BOL)*** results demonstrate that the SVM classifier with symbiotic organisms search(SOS)feature selection yields the best overall performance,with an area under the curve(AUC)of 83.27%and accuracy(ACC)of 83.33%for PV3D,and an AUC of 71.18%and ACC of 70.33%for *** demonstrates the stability and effectiveness of the proposed method in video-based deception detection tasks.