CAREER: Scalable Software and Hardware for Image and Video Processing Systems

点击翻译按钮获取中文摘要

This research involves improving quality, reducing cost, and shortening the time- to-market for image and video processing systems, such as video telephones, multimedia systems, and confocal microscopes. Image and video processing systems are difficult to design because they consist of a variety of algorithms implemented by a variety of electronic devices and software programs. In industry, it is common practice to divide the design of an image or video processing system into simpler parts and design the parts separately. Until all of the individual parts are designed and built, it is difficult to guarantee that the parts will work together, and know how the decisions in the design of an individual part will affect the cost and quality of the entire system. As an alternative, this research explores design techniques and computer tools that allow all the parts of an image and video processing system to be designed at the same time. The principal investigator is establishing a research and education program in the design and implementation of image and video processing systems. The research focuses on automating the design of scalable software and hardware for image and video processing systems. The performance of a scalable implementation improves as more processors are allocated to the task, until the number of processors reaches upper limit. Scalable hardware can expand to a variable number of processors, whereas scalable software adapts to the processors that are available. Scalable software is available on desktop computers, and scalable hardware is commonly used in high-performance computing. The education program trains undergraduate and graduate students in state-of-the-art design methods and technologies. As part of this training, the investigator is introducing new courses in communication system design, image and video processing, and embedded software systems. By forging research partnerships with industry, the investigator is developing new stude nt design projects to solve system design bottlenecks in industry.