雜訊通道中低位元率視訊壓縮

Abstract

"視訊壓縮" 已成為多媒體通訊中之不可或缺的元素. 然而, 經過壓縮後 的資料, 很容易受到通道中雜訊的影響. 在此篇研究中, 我們模擬了 H.263 視訊編碼器, 並將壓縮後的資料輸入 Rayleigh 衰減通道模型中. 結果顯示: 在雜訊能量不高的情況下, 經由解碼後的影像, 仍會受到嚴重 的損害. 運用錯誤控制碼來保護經過壓縮後的資料, 是一種有效的方法. 另一方面, 我們也採用 "重疊式正交轉換" 對靜態影像 (JPEG) 及動態影 像(H.263) 進行處理. 模擬結果顯示: 採用 "重疊式正交轉換" 的資料比 採用 "離散式餘弦轉換" 的資料具有較佳的視覺效果. 重疊轉換的概念似 有潛力發展出更為強健的源編碼方式. Video compression is an indispensavle element in multimedia communicationssystems. However, the compressed data are very sensitive to channel errors.In this study, we simulate the H.263 video coder and pass the compresseddata through a simulated mobile communication channel. Preliminary resultsindicate that the decoded images are severely damages in this situationeven when the background noise is not very high. The error control codes,such as BCH codes, are very useful for protecting the compressed data. Onthe other hand, the lapped orthogonal transform (LOT) was proposed toprocess both the still image (JPEG) and the video sequence (H.263).Experimental results showed that the LOT-processed data produced bettervisual quality than DCT processed ones. Potentially, a more robust sourcecoding scheme can be developed using a lapped transform concept.