DHB: The Process of Collective Behavior: Validation of a Unified Mathematical Model Using Computer Vision Tracking

DHB：集体行为的过程：使用计算机视觉跟踪验证统一数学模型

• 批准号: 0729363
• 负责人: Richard Ruback
• 金额: $ 75万
• 依托单位: Pennsylvania State Univ University Park
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2007
• 资助国家: 美国
• 起止时间: 2007-09-15 至 2012-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0729363&HistoricalAwards=false
• 关键词: DHB; Process; Collective; Behavior; Validation

This interdisciplinary project teams a social psychologist, a mathematician, two computer scientists, and a sociologist in research that combines a mathematical model of collective behavior with software for image tracking, classification, simulation, and animation. The hypothesis to be tested is that behavior in a large crowd results primarily from social influence within and between small groups, in contrast to existing models, which operate at the level of the individual or of the entire collective. The model unifies these approaches and allows for comparative assessment by varying the importance given to individual factors versus properties of the entire collective versus the influence of small-group interaction. Validation of the model will be based on novel computer vision analyses of videos of crowds of varying nature, structure, and size. The theory and the tools to be developed by this project can assist personnel in law enforcement, emergency management, and event management to minimize violence, speed up evacuations, and reduce accidental injury. The automated tracking software will provide real-time information far beyond what human observers can accurately determine.

这个跨学科的项目团队由一名社会心理学家，一名数学家，两名计算机科学家和一名社会学家组成，他们将集体行为的数学模型与图像跟踪，分类，模拟和动画软件相结合。要检验的假设是，在一个大的人群中的行为主要是由社会影响的结果在小团体内部和之间，与现有的模式，在个人或整个集体的水平上运作。该模型统一了这些方法，并允许不同的重要性给予个人因素与整个集体的属性与小群体互动的影响进行比较评估。 该模型的验证将基于对不同性质、结构和规模的人群视频的新颖计算机视觉分析。本计画所开发的理论与工具，可协助执法、紧急事件管理与事件管理人员，以减少暴力事件发生、加速疏散、减少意外伤害。自动跟踪软件将提供远远超出人类观察者可以准确确定的实时信息。