Multiterminal (MT) video coding considers the problem of separate compression and joint decompression of multiple correlated video sources. Driven by a host of emerging applications (e.g., in network communications, distributed video sensor networks and camera arrays, and distributed inference), MT video coding has become a very active area of research in the past ten years.
This research addresses both the theory of MT source coding and practice of MT video coding, with the aim of using the theory to guide practice. It involves: 1) identifying new classes of problems of quadratic Gaussian MT source coding with tight sum-rate bound or giving sufficient conditions for sum-rate tightness; 2) taking an approximation approach when sum-rate tightness cannot be proved; 3) performing MT source code constructions based on Slepian-Wolf coded quantization for random coding and hybrid random-structured coding; and 4) spearheading the practical application of MT video coding for camera arrays and distributed video sensor networks, with focus on depth camera assisted MT video coding.
