CAREER: Models and Representations for Digital Images and Video

Polytechnic University Of New York, Brooklyn NY

Investigators

Abstract

Increased connectivity to information networks and the widespread availability of high bandwidth links are expected to result in much increased interest in digital image and video applications providing very high quality and advanced processing and compression options. This research is geared toward addressing this need by providing more accurate representations, modeling, processing and compression for digital images and video. The main objective is the development of localized, multiresolutional models and the design of related estimation, optimization, and processing strategies. The developed techniques are specifically targeted at digital images and video depicting natural scenes, which are characterized by localized singularities separating regions of uniform smoothness/texture in the camera plane. Statistical multiresolutional models for the singularities combined with global optimization techniques constitute the foundations of the research track. The developed techniques are utilized in a variety of applications including compression, estimation of missing data, noise removal, segmentation and resolution enhancement. The results of this research are expected to have an impact in a variety of areas, such as wireless communications and multimedia applications over the Internet. Modern engineering education involves many courses that assume students are familiar with detailed technical background material. The educational component of this is geared toward addressing significant variations in student background by allowing the students to enhance their learning experience at their own pace. Educational objectives include the development of computerized, multimedia assisted course material, including problems and example solutions with rich annotations. Extensions of this material to distance learners, including outreach to broader communities of learners over information networks, are expected.