Collaborative Research: Towards Real-time, High throughput Insect Behavior Analysis

协作研究：实现实时、高通量昆虫行为分析

• 批准号: 1045239
• 负责人: Min Shin
• 金额: $ 22.38万
• 依托单位: University of North Carolina at Charlotte
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2010
• 资助国家: 美国
• 起止时间: 2010-09-15 至 2013-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1045239&HistoricalAwards=false
• 关键词: Collaborative; Research; Towards; Real; time

This project will develop an efficient method of analyzing a large quantity of animal behavior data collected in video. Animal behavior research is often limited by the large time investments needed to observe animal activities and extract quantitative information from this. Automating this process, in other words enabling a computer system to extract relevant information from videos of animal behavior, would result in radical time-saving for such studies. Not only that, but a system that would allow real-time analysis could enable researchers to change experiments mid-stream, as measurements of the behavior of observed individuals would become available immediately. This will allow a new level of control, and enable totally new experimental approaches. A tool will be developed to enable high-throughput, real-time analysis of insect behavior on a Desktop PC by fusing sophisticated image analysis software and powerful, rapidly developing Graphical Processor Unit (GPU) hardware technology. The tool will be affordable (about $10k/unit), easy-to-maintain (it's a PC) while providing a real-time, automated analysis of insect behavior increasing the throughput by up to 400 times. Ant colonies, one of the model organisms for which we will validate the tool, are uniquely suitable to demonstrate both sophistication and complexity of biological systems in outreach. In addition, graduate students will be provided interdisciplinary training through the project. Interdisciplinary research, in particular computer science-biology interactions, is the future, and this type of project will serve as a great training mechanism for the students. Most importantly, the developed tool will be disseminated to other researchers by making the software available for download through the project website.

该项目将开发一种有效的方法来分析视频中收集的大量动物行为数据。 动物行为研究通常受到观察动物活动并从中提取定量信息所需的大量时间投资的限制。 自动化这一过程，换句话说，使计算机系统能够从动物行为的视频中提取相关信息，将大大节省此类研究的时间。不仅如此，一个允许实时分析的系统可以使研究人员在中途改变实验，因为对被观察个体行为的测量将立即可用。这将允许新的控制水平，并使全新的实验方法成为可能。将开发一种工具，通过融合先进的图像分析软件和功能强大、快速发展的图形处理器单元（GPU）硬件技术，在台式PC上实现对昆虫行为的高通量、实时分析。该工具将是负担得起的（约1万美元/台），易于维护（它是一台PC），同时提供实时，自动化的昆虫行为分析，增加高达400倍的吞吐量。 蚁群是我们将验证该工具的模型生物之一，它是唯一适合展示生物系统复杂性和复杂性的模型。此外，研究生将通过该项目获得跨学科培训。跨学科研究，特别是计算机科学-生物学相互作用，是未来，这种类型的项目将作为一个伟大的培训机制的学生。最重要的是，将通过项目网站提供软件供下载，向其他研究人员传播所开发的工具。