Progressive Side Information Refinement with Non-Local Means Denoising in Distributed Video Coding

Wyner-Ziv (或簡稱WZ) 視訊編碼為分散式視訊編碼 (或簡稱DVC) 中一種特別的實作方式，它是基於Slepian-Wolf與Wyner-Ziv 的理論；相較於傳統以預測移動為基礎移動的視訊編碼器，這種新的視訊壓縮架構是在解壓縮端進行資料時間域上關聯性的利用而非在壓縮端。至今，目前最先進的WZ編碼器，其壓縮效率跟目前最先進以預測移動為基礎的編碼器比較起來還是有段差距，尤其是在視訊內容移動較為複雜的情況下。這主要是因為旁資訊 (或簡稱SI) 品質的低弱，在眾多熱門的WZ壓縮器中，旁資訊被視為在解壓縮端產生的原始畫面預測。這篇論文提出一個轉換域W編碼器的循序旁資訊改善技術，基於去雜訊的概念並且整合至最先進的WZ編碼器的解壓縮流程。藉由非區域平均去雜訊演算法 (或簡稱NLM) 同時探索空間與時間上的相似性，旁資訊在解碼的過程中被循序的改善。實驗結果顯示，在移動快速以及遠的畫面群集 (或簡稱GOP) 的影片中，編碼器壓縮效率可以提升多達2.6分貝相較於同一個編碼器但未採用本篇提出的技術。Wyner–Ziv (WZ) video coding is a particular case of distributed video coding (DVC), the recent video coding paradigm based on the Slepian–Wolf and Wyner–Ziv theorems which exploits the source temporal correlation at the decoder and not at the encoder as in conventional predictive-based video coding. Until now, the coding efficiency of state-of-the-art WZ codecs is still far from that of the state-of-the-art prediction-based codecs, especially for high and complex motion contents, especially for the high and complex motion contents. It is mainly due to the poor quality of the side information (SI), which is an estimate of the original frame created at the decoder in the most popular WZ video codecs. This thesis proposes a progressive side information refinement scheme of transform domain WZ video coding, which is based on the denoising concept and integrated into the decoding process of a popular WZ video codec. The side information is progressively improved by explore both temporal and spatial similarity using non-local means (NLM) denoising algorithm as the decoding proceeds. The results obtained reveal gains up to 2.6dB in the RD performance against the performance for the same codec without the proposed technique for high motion sequences and long group of pictures (GOP) sizes