中国科学院行为科学重点实验室知识库

隐藏情绪识别对公共安全防范与预警具有重要的意义。微表情是揭示隐藏情绪的一条重要通道。但目前隐藏情绪研究较少且微表情因其细微幅度与快速出现等特性难以识别,其研究尚未在实际中广泛应用。因为,隐藏情绪的认知与表达机理亟需系统的研究,采集实际场景中的微表情数据,并以脑电信号辅助微表情的精确标注是提高微表情标注效率的有效途径。深入研究微表情识别方法,并辅以人脸颜色、注视估计和非接触生理信号等多通道数据,以检测与识别隐藏情绪。社会公共安全是隐藏情绪分析和识别的典型场景。面向精神疾病患者两害行为(即危害自身或他人的危险行为)风险评估和服刑人员会见场景隐藏情绪检测,可以有效地对相应系统和方法进行验证和修正。

It is of great value to recognize concealed emotions for early warning of public security issues. Micro-expression is a vital channel to reveal concealed emotions. However, there are relatively few studies on concealed emotions, and micro-expressions are challenging to recognize because of their subtle magnitude and short duration. Existing Laboratory studies of micro-expression have few practical applications. Therefore, the perception and expression of concealed emotion should be systematically investigated by collecting micro-expression samples in an ecological situation, while synchronically collecting EEG signals for better labeling of micro-expressions. We spot and recognize concealed emotions mainly through micro-expressions, accompanied by face color analysis, gaze estimation, and contactless physiological signals measurement. Then, we verify and modify our system and method in two realistic public security related application scenarios: a Recognition Assistant System for the aggressive and suicidal behaviors of psychiatric patients and a Concealed Emotion Detection System for prisoners interview.

国家自然科学基金项目（U19B2032,61772511）;; 社会安全风险感知与防控大数据应用国家工程实验室主任基金项目（18112403）