透過您的圖書館登入
IP:18.189.180.244
  • 學位論文

具雙重旁資訊產生機制之分散式視訊編碼系統

Distributed Video Coding with Dual Side Information Generation

指導教授 : 李昌明
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


隨著科技不斷發展,使用者對於多媒體的需求也日漸增加且大不相同,部分系統要求成本低廉的編碼硬體設計、低耗電量以及資料傳輸即時性,為滿足此類需求,分散式訊源編碼系統 ( Distributed Source Coding,簡稱DSC ) 就常被提出來解決此類問題,若將此概念應用於視訊處理,即稱為分散式視訊編碼系統 ( Distributed Video Coding,簡稱DVC )。 本篇論文將探討DVC系統中解碼端的旁資訊 ( Side Information,簡稱SI ) 產生機制,旁資訊對於Wyner-Ziv解碼端與影像重建的編碼效益非常重要。先前研究曾考量在TGOB架構之DVC系統中,討論到採用低複雜度旁資訊的產生機制並維持高效能的編碼效益,但在解碼端會因旁資訊的產生機制,而面臨編碼系統延遲的問題,因此無法充分滿足即時性視訊通訊的需求。透過對影像之原始資訊與旁資訊做相關性分析後,提出雙重旁資訊的編碼架構,透過修正區塊模式選擇的機制,進而調整TGOB的架構來滿足高編碼效率以及低延遲的視訊通訊,最後模擬各種雙重旁資訊系統的組合,其中Ave_Dec + Ave_Intra的方式與Bli_Intra方法的效能相比,最高可以節省6.26%的碼率。

並列摘要


As the rapid development of digital technology, users have different requirements in the multimedia applications. Some systems have constraint to design low-cost encoding module. In order to satisfy the requirement, distributed source coding (DSC) is discussed as one of solution. In the video system based on the DSC scheme (so called distributed video coding, DVC), the side information (SI) generator following the Wyner-Ziv (WZ) decoder and video reconstruction are essential to the coding efficiency. In the previous work in the DVC system with the temporal group-of-block (TGOB) structure, the low-complexity SI generations are discussed and some of them have gains in the rate-distortion performance. However, the coding latency in this codec is impractical for the real-time video communication. After analysis of data dependence in the TGOB structure, all possible SI generations with low computational complexity are evaluated. According to the coding dependence, the block mode decision in the encoder is modified. Furthermore, the dual-SI scheme is proposed to improve the LDPCA decoding and reconstruction. Therefore, high-efficiency and low-latency video communication is available. Besides, the complexity (computation and memory) of the SI generation can be reduced in advance. In summary, the rate-reduction can be up to 6.26% in the dual-SI scheme with Ave_Dec + Ave_Intra, compared to the bi-linear interpolation intra (Bli_Intra) method.

參考文獻


[1] D.C. Tsai, C.M. Lee, and W.N. Lie, “Dynamic key block decision with spatio-temporal analysis for Wyner-Ziv video codec,” Proc. of IEEE Int’l Conf. on Image Processing (ICIP 2007), Vol. 6, pp. VI-425 - VI-428, November 2007.
[2] D. Slepian, and J. K. Wolf, “Noiseless coding of correlated information sources,” IEEE Trans. on Information Theory, Vol. 19, No. 4, pp. 471-480, July 1973.
[3] D. Wyner, and J. Ziv, “The rate-distortion function for source coding with side information at the decoder,” IEEE Trans. on Information Theory, Vol. 22, No.1, pp. 1-10, January 1976.
[5] S.Y. Chien, T.Y. Cheng, S.H. Ou, C.C. Chiu, C.H. Lee, V.S. Somayazulu, and Y.K. Chen, “Power comsumption analysis for distributed video sensors in machine-to machine networks,” IEEE Trans. on Emerging and Selected Topiocs in Circuits and Systems for Video Technology, vol. 3, no. 1, pp. 55-64, March 2013.
[10] R. Puri, A. Majumdar, and K. Ramchandran, “PRISM: A video coding paradigm with motion estimation at the decoder,” IEEE Trans. on Image Processing, Vol. 16, No. 10, pp. 2436-2448, October 2007.

延伸閱讀