FACIAL EXPRESSION ANALYSIS BY COMPUTER IMAGE PROCESSING

面部表情传达关于情绪的信息，调节 人际行为和人的知觉，指标大脑功能， 对于评估先兆婴儿来说也是必不可少的。当前的人类观察者 然而，面部表情分析的方法是劳动密集型和 很难跨实验室和随着时间的推移实现标准化。这些因素 迫使调查人员使用效度收敛的不太具体的系统 通常是未知的。使计量更加严谨、量化、可行 面部表情在不同的应用中，我们的跨学科 研究小组，在面部表情分析和 计算机化的图像处理，将发展面部的自动化方法 表情分析。 面部表情将被视频记录在定向的面部 行动任务和情感小插曲。对于每个受试者，一个3-D电子设备 线框人脸模型将通过拟合数字化图像 将受试者的中性脸部移至基本模型。计算机化特征检测 跟踪程序将使用该人脸模型来分析输入视频 图像序列。神经模式识别算法将测量和 基于测量结果对面部表情动作单元进行分类 特征跟踪算法。用户界面将允许用户定义 面部配置(根据EMFAC、MAX或其自己的规格)和 生成时间序列或汇总数据文件，以便在SPSS中进行即时分析 或其他统计软件。 面部表情分析的自动化方法将不再需要 用于人类观察者对面部表情进行编码(大大减少了编码 时间和人员成本)，促进标准化测量，使之成为可能 收集和处理更大、更具代表性的数据集，以及 开辟了研究和临床应用的新领域。例如, 自动化方法和人工观察者方法之间的比较可以告诉我们 人们处理情绪表达。与需要熟练技术的方法不同 观察到，自动化方法将很容易从 用于诊断和患者分析的实验室和临床 通讯。

Facial expression communicates information about emotions, regulates interpersonal behavior and person perception, indexes brain functioning, and is essential to evaluating preverbal infants. Current human-observer methods of facial expression analysis, however, are labor intensive and difficult to standardize across laboratories and over time. These factors force investigators to use less specific systems whose convergent validity is often unknown. To make feasible more rigorous, quantitative measurement of facial expression in diverse applications, our interdisciplinary research group, with expertise in facial expression analysis and computerized image processing, will develop automated methods of facial expression analysis. Facial expressions of emotion will be videorecorded in a directed facial action task and emotion vignettes. For each subject, a 3-D electronic wireframe face model will be created by fitting a digitized image of the subject's neutral face to a basic model. Computerized feature detection and tracking procedures will use this face model in analyzing input video image sequences. Neural pattern recognition algorithms will measure and classify facial expression action units based on measurements from the feature tracking algorithms. A user interface will permit users to define facial configurations (per EMFACS, MAX, or their own specifications) and generate time series or summary data files for immediate analysis in SPSS or other statistical software. The automated method of facial expression analysis will eliminate the need for human observers in coding facial expression (greatly reducing coding time and personnel costs), promote standardized measurement, make possible the collection and processing of larger, more representative data sets, and open new areas of investigation and clinical application. For example, comparisons between automated and human-observer methods can inform how people process emotion expressions. Unlike methods requiring skilled observation, the automated method will be readily transferable from the laboratory tot he clinic for use in diagnostics and in analysis of patient communications.