Slepian-Wolf SPIHT video coding scheme without feedback

Abstract

The proposed scheme is an extension from Set Partitioning In Hierarchical Trees (SPIHT) algorithm. It combines the SPIHT algorithm with Slepian-Wolf (SW) theorem. In most of the video coding scheme that adapted the Slepian-Wolf or Wyner-Ziv (WZ) theorem, a feedback channel is usually incorporated to correct errors that occurred at the decoder side. The used of feedback is a big drawback in system where latency or memory constraints are a critical issue. In order to investigate the possibility in tolerating the errors, and achieve a reasonable result, the feedback channel is not adapted in the proposed scheme. In the proposed scheme, the video sequence will be divided into two types, which defined as Key frame and SW frame. The Key frame is intracoded with SPIHT algorithm alone, while the SW frame is coded using the proposed algorithm. First of all, the SW frame is coded using the SPIHT algorithm to extract the Significant bits, Sign bits, and Refinement bits. Then, the Refinement bits will be further coded by a Coset encoder. The results show that the proposed scheme is perform slightly better than using SPIHT algorithm alone.