【作者】 徐辉；

【导师】 张智君；

【作者基本信息】 浙江大学 ， 应用心理学， 2010， 硕士

【摘要】 面部表情是社交信息的重要来源,但作为面部表情识别的核心问题之一——表情识别的加工机制,还未有一致性结论。目前有三种具有代表性的理论,分别是结构说、特征说和整体说。对人脸图像的空间频率进行滤波,是研究表情识别加工过程的一个有效方法。本研究采用空间频率滤波技术,对表情图片进行高、低通滤波处理,考察不同空间频率对表情识别的影响,并进而探讨结构和特征信息对表情识别的影响。本研究包括三个实验：实验一在正常、低空间频率和高空间频率三种情境下呈现不同的表情,让被试进行判断,以此考察高、低空间频率对表情识别的影响。实验二在低空间频率、“低空间频率+眼”以及“低空间频率+嘴”三种情境下呈现表情,要求被试进行判断,以此考察结构信息对表情识别的作用。实验三则研究在“低空间频率+眼”、“低空间频率+嘴”以及眼、嘴单独呈现的情境下,让被试判断表情,以此考察特征信息对表情识别的作用。本研究获得的以下主要结果：(1)不同表情的识别存在不同的加工机制。对“高兴”和“悲伤”表情的识别,符合“整体说”,而对“愤怒”和“恐惧”表情的识别,却与“特征说”一致。(2)对表情的识别存在“高兴优势效应”,且这种优势效应是由整体结构引起的。(3)对低频信息的加工比高频信息更有效。(4)在识别“高兴”表情时,嘴的作用比较重要,而在识别“愤怒”、“恐惧”和“悲伤”表情时,眼睛的作用则更重要。总之,本研究结论表明,不同表情的识别存在不同的加工机制,不同的特征信息在不同表情的识别中具有不同的重要性。更多还原

【Abstract】 Facial expressions are an important source of social information. However, there is not a consistent finding in facial expression processing mechanism which is a key aspect of facial expression recognition area. There are three representative theories, including configural model, featural model and holistic model. Meanwhile, analyzing spatial frequency component is a good way to study facial expression. Therefore, in the present study we study the influence of different spatial frequency on facial expression recognition and explore the impact of configural and featural information on facial expression recognition using high and low spatial frequency pictures.The present study includes three experiments:Experiment 1 aimed to investigate the influence of spatial frequency on facial expression recognition, in which participants were showed normal, low pass filter and high pass filter pictures, and were required to judge what expression they saw. The second experiment was to explore the impact of configural information on facial expression recognition by using the pictures of low pass filter, low pass filter+eye, low pass filter+mouth. The last experiment was further to study the influence of featural information on facial expression recognition through showing the pictures of low pass filter+eye, low pass filter+mouth, and eye and mouth only.The main findings of the present study are as below:(1) Facial expressions differ in processing. The process of happy and sad face is consistent with holistic model; the process of angry and fear face is identical with featural model.(2) There is "happy face advantage", and this effect is caused by holistic structure. (3) The processing of low-spatial-frequency information is more effective than the one of high-spatial-frequency information.(4) Mouth is important for identifying happy face, and eyes are crucial to recognizing anger, fear and sad face.To sum up, there is different processing mechanism in recognizing facial expressions. The importance of feature information differs in recognizing facial expressions.更多还原