CDI-Type I: Collaborative Research: A Bio-Inspired Approach to Recognition of Human Movements and their Styles

该奖项是根据2009年美国复苏和再投资法案(公法111-5)资助的。这项研究的目标是开发受生物启发的算法，以识别人类活动视频中的人体动作和动作风格。该方法基于腹侧视觉通路中静态三维(3D)形状结构的新表示，该结构由3D结构片段的配置组成。这个项目使用神经实验来验证一种基于4D(空间和时间)运动结构(SIM)片段的运动3D形状的类似的四维(4D)表示。这些SIM片段模型用于开发从人类活动的视频中自动提取SIM片段的算法。混合系统识别和聚类技术被用来学习用于识别的人体动作词典。这本词典反过来又会影响神经实验的设计。该项目的智能优势在于通过神经科学、计算机视觉、机器学习和动力学系统的计算和实验分析之间的交互，大大缩小了人和机器对人类运动的感知之间的差距。通过将分类、聚类和系统识别方法扩展到由混合动态模型集合生成的时间序列数据，该项目还有可能在机器学习和动态系统方面取得进展。潜在的更广泛影响包括在监视、安全、辅助家庭生活、婴儿护理、远程沉浸和运动员运动分析中的应用。该项目还赞助了一场比赛，由来自巴尔的摩的初中生和高中生制造小型机器人，并赞助了El Alliance项目，以留住研究生院代表不足的少数族裔。

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).The objective of this research is to develop bio-inspired algorithms for recognizing human movements and movement styles in videos of human activities. The approach is based on a new representation of static three-dimensional (3D) shape structure in the ventral visual pathway consisting of configurations of 3D structural fragments. This project uses neural experiments to validate an analogous four-dimensional (4D) representation of moving 3D shapes based on 4D (space and time) structure-in-motion (SiM) fragments. These SiM fragment models are used to develop algorithms for automatically extracting SiM fragments from videos of human activities. Hybrid system identification and clustering techniques are used to learn a dictionary of human movements used for recognition. This dictionary, in turn, influences the design of the neural experiments. The algorithms are evaluated using a real-time tele-immersion system.The intellectual merit of this project is to substantially reduce the gap between human and machine perception of human movements through the interaction between computational and experimental analyses from neuroscience, computer vision, machine learning, and dynamical systems. This project also has the potential to generate advances in machine learning and dynamical systems, by extending classification, clustering and system identification methods to time-series data generated by collections of hybrid dynamical models.Potential broader impacts include applications in surveillance, security, assisted home living, infant care, tele-immersion, and athlete motion analysis. The project also sponsors a competition where small robots are constructed by middle and high-school students from Baltimore, and the EL Alliance program for retaining underrepresented minorities at graduate schools.