Facial Expression Recognition is a newly developing research direction in the Field of artificial intelligence. The one of the study aim is to recognize Human Facial Expression automatically and analyze the emotion in artificial intelligence products, such as robots. Automatic Facial Expression Recognition is an important part of Computer Vision Research with vital value of research and application for further enhancing the friendly and intelligent human-computer interaction. Study on neural mechanism for facial expression, a very important point is based visual perception research. Early visual areas of primates are retinotopically organized, so that the visual field is mapped in each area along two orthogonal axes: polar angle and eccentricity. The center/periphery organization, that is, eccentricity mapping, is one of the most striking and robust organizational principles in the primate visual cortex. The center/periphery organization extends into higher-order visual areas, whereas the polar angle representation in these areas is cruder, and orderly representations of the visual field meridians are absent. Despite the evident importance of eccentricity maps, their possible relationship to object recognition has received little attention, and the possible effect of this organization on the way different object categories are represented in the human brain has not been studied. Such as facial expression occurs in the visual field direction or different sizes of human cognitive effect is also different. This is to introduce a facial expression recognition is closely related to the concept of the visual perception of the cognitive mechanism. Some researchers proved that the use of the central visual field stimuli of different facial expressions and different brain areas related to research. Such as the amygdalar processing the fearful facial expressions, anterior insula and putamen are correlated with disgust recognition of facial emotional expressions. Some researchers use ethological methods studies demonstrated that the peripheral vision on recognition of facial expressions of preference. But the recognition of facial expressions of neural mechanisms is still not clear conclusion in the peripheral visual cortex. The present study used fMRI and a novel wide-view visual presentation system aiming to explore the human ability to detect facial expressions in the peripheral visual field as a function of eccentricity. We chose to study two emotional facial expressions (neutral, fear and joy) and four view field. As fear is more indicative of imminent danger than disgust, requiring a rapid response. In order to test for the possibility of more efficient emotion detection in peripheral vision, a gender discrimination task was used as a control. Thus, we hypothesized that it would be better detection of emotion, especially fear, compared with gender in the peripheral visual field.
面部表情识别是人类特有的高级脑功能,其大脑机制至今尚未阐明。从空间视知觉的角度研究面部表情识别的大脑加工模式,可以很好地控制面部表情的强度和降低无关变量的影响,为研究大脑机制提供新视角。本项目将利用功能核磁共振技术(fMRI)和宽视野视觉图像显示设备,对被试呈现空间视觉(中心和周边)的面部表情图片(中性、高兴和恐惧),利用一元线性模型(GLM)比较不同表情在视觉皮层的激活区域及其信号变化,探讨视觉皮质对中心和周边的面部表情认知的神经机制差异。进一步利用多体素解析方法识别中心-周边面部刺激空间种类,分析视觉皮质对中心和周边面部表情认知的差别,确认中心和周边视觉皮层对面部表情识别的认知偏好,最终确定视觉皮质空间表情认知的神经模型。本研究将解决面部表情认知的科学难题,为面部表情识别机制解明、脑部图像再现、计算机视觉和仿真机器人研究提供理论依据。
面孔和面部表情识别是人类视知觉的高级脑功能,其神经机制至今不明。从空间视知觉的角度研究面部表情识别的大脑加工模式,可以很好地控制面部表情的强度和降低无关变量的影响。本项目利用功能核磁共振技术(fMRI)和宽视野视觉图像显示设备,对被试呈现空间视觉(中心和周边)的面孔识别进行了系列研究,利用一元线性模型(GLM)比较面孔和非面孔在视觉皮层的激活区域及其信号变化,探讨了视觉皮质对中心和周边的面孔的神经机制差异,并研究了中心和周边视觉皮层对面孔识别的认知偏好。得出在宽视野的条件下,侧视觉皮质区域对不同的刺激种类具有偏心角度的偏好性的结论;随后,我们将注意力集中在人脑对于人脸的面部表情的识别机制上,从行为学实验中得出对于所有表情来说,均有从中心到周边,正确率越来越低,反应时越来越长的特点;在周边视野,被试对于高兴表情的认知能力最高,对厌恶表情的认知能力最低的结论。.进一步本研究以正立或倒置的面孔和非面孔图片作为刺激设计面孔倒置效应实验,针对面孔特异性加工和构形加工的大脑机制进行了探究。为了对实验中涉及的特征指标进行进一步验证,本研究还设计了高兴-悲伤面部表情加工实验,以建立更加全面的脑机制模型。两组实验均采集了受试者的16个导联的脑电数据,基于独立分量分析对脑电数据进行预处理以提高信噪比,基于健康受试者脑电数据的事件相关电位、锁相能量和非锁相能量建立面孔加工的脑机制模型,并利用该模型对精神分裂症患者的面孔加工障碍的脑机制进行探究。最后,本研究还建立了基于事件相关电位-锁相能量-非锁相能量三重指标的面孔加工脑机制模型,选取N170成分的幅值和潜伏期作为ERP特征指标,选取θ频段、α频段和低频β频段的锁相能量作为锁相能量特征指标,选取γ频段的非锁相能量作为非锁相能量特征指标。研究结果显示,人们更倾向于对面孔类刺激采用构形加工,即面孔特异性加工。
{{i.achievement_title}}
数据更新时间:2023-05-31
涡度相关技术及其在陆地生态系统通量研究中的应用
基于 Kronecker 压缩感知的宽带 MIMO 雷达高分辨三维成像
环境类邻避设施对北京市住宅价格影响研究--以大型垃圾处理设施为例
基于SSVEP 直接脑控机器人方向和速度研究
居住环境多维剥夺的地理识别及类型划分——以郑州主城区为例
基于立体宽视野脑功能成像技术的跨通道情绪识别脑机制研究
抑郁症肝气郁证患者抑郁眼神、面部表情和姿态多模态识别及其脑区功能定位
社会权力影响面部表情加工的ERP研究
基于面部肌肉运动模型的三维人脸表情识别研究