具有情绪主动感知能力的人-机交互（HCI）系统，由于能够更好地理解人的意图，已成为一个新的研究热点。眼睛作为情绪表达的窗口，眼动信息将会为情绪识别提供新的重要线索。为了明确眼动信息在情感计算中的作用，构建基于眼动的情绪识别新方法，本项目将以多通道眼电信号（EOG）为研究对象，围绕其对情绪的表征与识别展开研究，具体包括：1）确定情绪EOG生成模型，研究眼动成份随时间与被试变化时情绪表征的稳定性；2）探索基于多导EOG信号的情绪空域表征方法；结合同步采集脑电与视频眼动信号的情绪分析结果，在时/频/空域上系统开展眼动模式与情绪状态间映射关系的研究；3）利用眼动单元上下文信息，建立“场景感知子模型”；联合“眼动单元识别子模型”实现具有场景感知能力的情绪识别新方法。项目研究内容面向自然HCI发展前沿，其研究成果将会对基于眼动信息的情绪计算提供理论基础和技术支撑，具有广阔的应用前景与学术价值。

Human-Computer Interaction (HCI) with the ability of emotional active perception has become a new research hotspot because it can understand user’s intention. Eye movement information will provide new significant clues for emotional recognition since eye can be considered as a window of emotional expression. In order to clarify the role of eye movement information in affective computing and build a new eye movement-based emotion recognition method, this project will focus on the study regarding emotion representation and recognition using multi-channel electro-oculography (EOG) signals. The main research contents are as follows: 1) Determine the emotional EOG generation model to acquire richer eye movement independent components, study their stability of emotional representation under different sections and subjects; 2) Explore the spatial representation method using multi-channel emotional EOG signals, research the reflection relationship between the eye movement pattern and the emotional state by combining the analysis of EEG and eye-movement video signals as well as their time / frequency / spatial characteristics; 3) Establish a “scene perception sub-model” using the contextual relationship among different eye movement units, achieve a new emotion recognition method with the ability of scene perception by fusing the output results of “eye movement unit recognition sub-model”. The research contents of this project are oriented to the frontier of natural HCI development. The research results will provide the theoretical foundation and technical support for the eye movement information-based effective computing, which have broad application prospects and academic values.