Advanced real-time algorithms for object-based video content extraction

用于基于对象的视频内容提取的高级实时算法

• 批准号: 262060-2008
• 负责人: Amer, Aishy
• 金额: $ 1.31万
• 依托单位: Concordia University
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2008
• 资助国家: 加拿大
• 起止时间: 2008-01-01 至 2009-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=395369
• 关键词: Advanced; real; time; algorithms; object

Video is becoming increasingly used in applications as diverse as entertainment, security, education, medicine, and even wireless devices. The introduction of video over mobile devices will accelerate this trend. For example, statistics Canada revealed that Canadians spent 21.6 hours per week watching TV in 2002, and security studies show that security concerns of government agencies, enterprises and individuals are driving the video surveillance market to a constant and robust expansion to be sustained for decades. A fundamental task in all these applications is the extraction of dynamic video content such as moving objects and their features. Video content is subject to different interpretation and video description can vary according to observers and applications. Video usually includes multiple objects, occlusion, and artifacts. The stable extraction of video content in real-world applications is, therefore, a challenging task.

视频越来越多地应用于娱乐、安全、教育、医疗甚至无线设备等各种应用中。移动设备上视频的引入将加速这一趋势。例如，加拿大统计局透露，2002年加拿大人每周花21.6小时看电视，安全研究表明，政府机构、企业和个人的安全担忧正在推动视频监控市场持续数十年的持续强劲扩张。在所有这些应用程序中的一个基本任务是提取动态视频内容，如移动对象及其特征。视频内容有不同的解释，视频描述可以根据观察者和应用而变化。视频通常包括多个对象、遮挡和伪影。因此，在实际应用中稳定地提取视频内容是一项具有挑战性的任务。