Attentional-Perceptual Scanpath for Emotion Recognition: A New Framework

用于情绪识别的注意感知扫描路径：一个新框架

• 批准号: RGPIN-2020-05575
• 负责人: RoyCharland, Annie
• 金额: $ 2.4万
• 依托单位: Université de Moncton
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2020
• 资助国家: 加拿大
• 起止时间: 2020-01-01 至 2021-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=717300
• 关键词: Attentional; Perceptual; Scanpath; Emotion; Recognition

The face has a functional value transmitting information about our emotional state. Recognition of emotional facial expressions is crucial, allowing us to adapt our behaviour accordingly. The current program is centered on the processing and recognition of emotional facial expressions. The recognition rates for the expressions of the basic emotions such as happiness, anger, disgust, fear, surprise and sadness are high, surpassing chance level. However, not all emotions are recognized equally well. Variations are observed between emotions with happiness being the most accurately recognized, followed by anger, surprise, disgust, sadness, and finally, fear. While the pattern of errors and general recognition rates seem clear, the underlying processes are not well understood. The main goal of the current program is to test a new comprehensive framework called the Attentional-Perceptual Scan-path for Emotion Recognition (APSER) that I have developed. According to APSER, recognition of emotional facial expressions is accomplished by sequentially attending to individual features of the facial expression. Individual features are integrated into a unified representation and new features are processed until a label is selected. The first axis of the research program includes experiments to explore attentional-perceptual mechanisms. In effect, while the role of attentional-perceptual mechanisms in the recognition of emotional facial expressions is undisputable, it remains unclear how they operate. The second axis will test the new assumptions of APSER, which stipulate that recognition is based on the sequential attention to and perception of individual features of the facial expression as a function of the emotion expressed until the accurate label is selected. The third axis of the program focuses on the applicability of APSER from a developmental perspective as research has shown that children process emotional facial expressions differently than adults. The understanding of basic mechanisms in the recognition and processing of non-verbal behaviour are crucial pillars for a better understanding of social interactions and interventions.

脸具有传递我们情绪状态信息的功能价值。面部表情的识别是至关重要的，它能让我们相应地调整自己的行为。目前的程序主要集中在处理和识别情绪面部表情。对快乐、愤怒、厌恶、恐惧、惊讶、悲伤等基本情绪表情的识别率高于随机水平。然而，并不是所有的情绪都能被同样地识别。人们观察到情绪之间的变化，最能准确识别的是快乐，其次是愤怒、惊讶、厌恶、悲伤，最后是恐惧。虽然错误的模式和一般识别率似乎很清楚，但潜在的过程并没有得到很好的理解。当前项目的主要目标是测试我开发的一个新的综合框架，称为情绪识别的注意-知觉扫描路径（APSER）。根据APSER，对面部表情的情感识别是通过依次关注面部表情的各个特征来完成的。单个特征被集成到一个统一的表示中，新的特征被处理，直到选择一个标签。研究计划的第一个轴包括探索注意-知觉机制的实验。实际上，虽然注意知觉机制在识别情绪面部表情中的作用是无可争议的，但它们是如何运作的仍不清楚。第二个轴将测试APSER的新假设，该假设规定，识别是基于对面部表情的单个特征的连续关注和感知，作为所表达的情绪的函数，直到选择准确的标签。该计划的第三个轴侧重于从发展的角度来看APSER的适用性，因为研究表明儿童处理情绪性面部表情的方式与成人不同。理解非语言行为的识别和加工的基本机制是更好地理解社会互动和干预的关键支柱。