以壓縮感測為基礎之旁消息增強分散式視訊編碼

Abstract

隨著無線網路技術的進步與手持式行動裝置的普及，隨時隨地在任何平台上進行影像錄製、視訊播放、影像處理與影片分享已不再只是一個幻想。然而，要讓使用者即使在低功率、低運算能力的裝置上都能快速的取得、播放與分享數位影片，還有許多技術上的挑戰需要克服，例如降低感測與編碼計算所需之資源、提高壓縮的效率等等。近年來，分散式視訊編碼 (distributed video coding) 興起提供了影像壓縮新的方法。它顛覆了傳統的壓縮方式，將原本必須在編碼端大量計算的複雜度移至解碼端來做。另外，壓縮感測 (compressive sensing) 的技術，提供了將信號的壓縮與感測在同一時間完成的方法，大幅降低感測所需的成本。 在此篇論文中，我們以分散式視訊編碼結合壓縮感測技術，來降低感測與編碼計算所需之資源，並提高壓縮的效率。我們的系統以壓縮感測為基礎，探索影片空間上和時間上 (spatio-temporal) 的統計特性，做為解碼時輔助的旁消息。我們提出了藉由分析訊號的統計特性來增強旁消息 (side information) 品質的演算法，並利用旁消息計算出訊號的機率模型，做為可信度傳遞 (belief propagation) 解碼時使用的事前機率。實驗結果顯示，和現存其他以壓縮感測為基礎的方法相比，我們的方法大幅提升了訊號重建的結果。

In this thesis, a novel distributed video coding (DVC) scheme on the basis of compressive sensing (CS) that achieves low-complexity for encoding and efficient signal sensing is presented. Most CS recovery algorithms rely only on the signal sparsity. Yet, under DVC architecture, additional statistical characterization of the signal is available, which offers the possibility of achieving more precise CS recovery. First, a set of random measurements are acquired and transmitted to the decoder. The decoder then exploits the statistical characterization of the signal and generates the side information (SI). Finally, utilizing the SI, a Bayesian inference using belief propagation (BP) decoding is performed for signal recovery. The proposed CS-DVC system offers a more direct way of signal acquisition and the potential for more precise estimation of the signal from random measurements. Experimental results indicate that the generated SI can improve the signal reconstruction quality in comparison with a CS recovery algorithm which relies only on the signal sparsity.