This paper presents the FACSHuman software program, a tool for creating facial expression materials (pictures and videos) based on the Facial Action Coding System (FACS) developed by Ekman et al. (2002). FACSHuman allows almost all the Action Units (AUs) described in the FACS Manual to be manipulated through a three-dimensional modeling software interface. Four experiments were conducted to evaluate facial expressions of emotion generated by the software and their theoretical efficiency regarding the FACS. The first study (a categorization task of facial emotions such as happiness, anger, etc.) showed that 85% of generated pictures of emotional expressions were correctly categorized. The second study showed that only 82% of the most-used AUs were correctly matched. In the third experiment, two independent FACS coders rated 47 AUs generated by FACSHuman using the standard methodology used in this kind of task (AU identification). Results showed good-to-excellent codification rates (64% and 85%). In the last experiment, 54 combinations of AU were evaluated by the same FACS coders. Results showed good-to-excellent codification rates (68–82%). Results suggested that FACSHuman could be used as experimental material for research into nonverbal communication and emotional expression.

本文介绍了 FACSHuman 软件程序，这是一种基于 Ekman 等人开发的面部动作编码系统 (FACS) 创建面部表情材料（图片和视频）的工具。(2002)。FACSHuman 允许通过三维建模软件界面操作 FACS 手册中描述的几乎所有动作单元 (AU)。进行了四个实验来评估软件生成的情绪面部表情及其关于 FACS 的理论效率。第一项研究（幸福、愤怒等面部情绪的分类任务）表明，85% 的生成的情绪表情图片被正确分类。第二项研究表明，只有 82% 最常用的 AU 被正确匹配。在第三个实验中，两个独立的 FACS 编码器对 FACSHuman 使用此类任务（AU 识别）中使用的标准方法生成的 47 个 AU 进行评级。结果显示良好到优秀的编码率（64% 和 85%）。在上一个实验中，由相同的 FACS 编码器评估了 54 种 AU 组合。结果显示良好到优秀的编码率（68-82%）。结果表明，FACSHuman 可用作研究非语言交流和情感表达的实验材料。