NYI: Research in Physics-Based Computer Vision

PI正在开发一个多传感器环境，用于精确、校准的视觉实验，控制成像光学和传感器参数，以促进对场景亮度、颜色和照度等因素的快速、准确测量。该环境用于开发和测试基于各种物理模型的一系列视觉传感器和恢复算法。最终目标是一个基于单一计算框架的综合视觉系统，具有增强的能力，这是由于纳入了有关物理原理的信息。该中心还正在为计算机视觉的本科生和研究生教育开发一项全面的课程，其中包括一个图像库、用于概念演示的研究型视觉辅助工具，以及用于处理真实图像的动手体验的互动软件。

The PI is developing a multi-sensor environment for precise, calibrated vision experiments, controlling imaging optics and sensor parameters to facilitate quick, accurate measurement of factors such as scene brightness, color, and illlumination. This environment is used to develop and test a series of vision sensors and recovery algorithms based on a variety of physical models. The ultimate goal is an integrated vision system based on a single computational framework for general visual perception, with enhanced capabilities due to the incorporation of information about physical principles. The PI is also developing a comprehensive curriculum for undergraduate and graduate education in computer vision, incorporating an image library, research-based visual aids for concept demonstration, and interactive software for hands-on experience in processing real images.