適應性H.264/SVC空間層際可調式視訊編碼之快速模式決策機制

H.264/AVC視訊編碼標準已廣泛地運用於各種場合。雖然H.264/AVC可以在視訊品質與位元率之間取得良好的平衡，然由於不同應用的需求與環境的限制，經由H.264/AVC所產生之視訊串流並不一定適合於各種應用場合的頻寬下使用。為了適應不同的環境需求與傳輸限制（例如有限的網路頻寬、有限的接收端解析度等），具備可調式編碼能力的H.264/SVC乃應運而生。H.264/SVC架構之特性在於一次編碼下可產生出不同品質的視訊子串流以提供不同環境所需。傳送端可依據不同的環境需求將不同品質的H.264視訊串流提供給接收端進行解碼。H.264/SVC所提供的可調性可概分為空間（Spatial）、時間（Temporal）與品質（Quality）可調性，編碼端可依據不同的環境條件進行不同的可調性編碼。由於H.264/SVC的編碼端可一次產出各種不同品質的視訊串流，因此其編碼端的運算負擔遂相當地沈重。此外經由統計，我們發現移動估測（Motion Estimation；ME）的運算複雜度佔整體編碼運算的80%上，而空間層越高時，因為畫面尺吋越大，故隨之而來的運算複雜度也越高。為能加速H.264/SVC的編碼效能，本論文將研究重點放在建立空間可調性的快速編碼決策機制。利用基礎層（Base Layer；BL）與增強層（Enhancement Layer；EL）的對應關係性，我們可以有效地將增強層候選模式加以限縮。此外透過大量地統計基礎層移動向量強度與增強層的模式對應關係，我們可以建立以基礎層移動向量強度來決策增強層模式的機制。經由實驗證明，所提出之演算法可以有效地提升增強層的編碼效能，相較於 JSVM9.18，所提出之演算法在編碼效能上平均約可以達到 84.14%的編碼時間節省，並保持良好的視訊品質與位元率。

關鍵字

可調式視訊編碼；快速模式決策；空間可調式層際預測；移動向量強度

並列摘要

H.264/AVC has been successfully applied in many applications. Although a very good trade-off between visual quality and bit-rate expense can be obtained by using H.264/AVC, the generated bit stream is not applicable to all the applications due to limited bandwidth available or processing capability of decoder. For this reason, the H.264/SVC that is capable of generating bit-stream with different visual quality has been developed. In H.264/SVC, bit streams with different quality can be generated in one time during the coding process. The SVC provides flexible spatial, temporal, and quality setting to adapt itself to different coding requirements. Based on our observation, we find that the process of motion estimation takes more than 80% of the computational complexity during the coding process. In addition, more computations are required for higher enhancement layer. To alleviate the computational burden of the encoder, we focus on the fast mode decision of enhancement layer when spatial scalability is considered. By using the relationship between base layer and enhancement layer, the most probable prediction mode of enhancement layer can be confined to a small subset. In addition, we also find a relationship between the motion vector strength of the collocated macroblock in the base layer and corresponding macroblocks in the enhancement layer so that the prediction mode in the enhancement layer can be determined in a more efficient manner. As we can see in the experiment, an average time-saving of 84.14% can be obtained by using the proposed approach with only a minor degradation on the bit-rate and PSNR performance when compared with JSVM9.18 reference software.

並列關鍵字

H.264/SVC ； Scalable video coding ； Fast mode decision ； Spatial scalability for inter-layer prediction ； Motion vector strength

參考文獻

[2] T. Wiegand, G. Sullivan, G. Bjontegaard, and A. Luthra, “Overview of the H.264/AVC video coding standard,” IEEE Trans. Circuits Syst. Video Techn,VOL.13,NO.7,JULY 2003

[4] H. Schwarz, D. Marpe, and T. Wiegand, “Overview of the scalable video coding extension of the H.264/AVC standard,” IEEE Trans. Circuits Syst. Video Technol., pp. 1103–1120, vol. 17, No. 9 , Sep. 2007.

[5] X. Lu, and G.R. Martin, “Fast mode decision algorithm for the H.264/AVC scalable video coding extension,” IEEE Transactions on Circuits and Systems for Video Technology, pp. 846-855, vol. 23,No. 5, May 2013.

[6] H.Li, Z.Li and C.Wen, “Fast mode decision algorithm for inter-frame coding in fully scalable video coding,”IEEE Transaction on Circuit and Systems for Video Technology, pp.889- 895, vol. 16, No 7, July 2006.

[7] G. Goh, J. Kang, M. Cho, and K. Chung, “Fast mode decision for scalable video coding based on neighboring macroblock analysis,” in Proc. ACM Symp. Appl. Comput. , pp. 1845–1846., 2009

國際替代計量

適應性H.264/SVC空間層際可調式視訊編碼之快速模式決策機制

未授權

主題瀏覽