Reliable and efficient image/video transmission presents an extreme challenge in harsh mobile environments. In this thesis, we develop a unique set of techniques, including combined source and channel coding (CSCC) schemes, bursty error countermeasures, and pre/post-processing techniques, to support mobile image/video transmission with substantially improved error resilience capability and transmission efficiency....[ Read more ]

Reliable and efficient image/video transmission presents an extreme challenge in harsh mobile environments. In this thesis, we develop a unique set of techniques, including combined source and channel coding (CSCC) schemes, bursty error countermeasures, and pre/post-processing techniques, to support mobile image/video transmission with substantially improved error resilience capability and transmission efficiency.

Unlike previous work where the bursty nature of errors was often ignored, the proposed CSCC takes both the correlated source characteristics and bursty error structure into account. To study the error structure in wireless channels, a finite state Markov model (FSMM) is developed for representing a correlated fading channel with diversity. A recursive algorithm, based on this model, is then proposed for an efficient estimation of block codes performance, so as to find the best error detection/correction coding design in fading environments. By incorporating the error sensitivities of image/video data, a unique methodology for optimization of CSCC schemes is then obtained.

In addition to CSCC, a low complexity spatial concealment, which is based on multi-directional interpolation, is developed in conjunction with block-shuffling, in order to effectively recover erroneous blocks in image transmission. Further, an H.263-based error resilient video transmission system is proposed with powerful bursty error countermeasures achieved by the use of diversity and CSCC. Robust decoding together with a combined spatial-temporal concealment is employed to mitigate the effects of decoding mis-match caused by residual errors. Meanwhile, an efficient motion estimation algorithm is devised in the encoder as another enabling technique to support low complexity video codec implementation for wireless applications.

For an integrated multimedia system design, we also investigate the performance of combined modulation and diversity in terms of enhanced cellular spectrum efficiency. Useful results are presented, and tradeoffs among spectrum efficiency, power efficiency and implementation complexity are studied.