影像封包遺失是在不穩定的網路進行影像傳輸所無法避免的錯誤。在H.264中，若封包遺失發生在I-frame更是會造成嚴重的錯誤累積，導致影像品質大幅下降。H.264在處理這類問題時，提供了編碼端的Flexible Macroblock Ordering(FMO)以及解碼端的Error concealment機制來解決。本篇論文將探討I-frame上的空間域錯誤藏匿，並將影像內容分為三種:平滑，邊緣，以及複雜性貼圖，對應的錯誤藏匿演算法分別是:二元內插，方向性內插，最佳鄰近貼圖。在編碼端，我們利用FMO的disperse分群模式以確保四個鄰近區塊的非同時遺失，並藏入各鄰近區塊的分類資訊。在解碼端，我們利用取出的分類資訊來針對不同特性的遺失區塊進行錯誤藏匿。我們特別對方向性內插以及最佳鄰近貼圖演算法作一些改進使其達到更好的藏匿效果。實驗結果顯示，我們的方法在視覺上改善了許多。

Packet loss is an inevitable problem on video transmission in unsteady network environment. A packet loss occurs in the I-frame of H.264 video causes a big degradation over video quality. H.264 supports two schemes to deal with this kind of troubles: Flexible Macroblock Ordering (FMO) at the encoder and error concealment (EC) at the decoder. In this paper, we focus on the Spatial Error Concealment (SEC) on the I-frame and classify a macroblock into a smooth block, edge block or texture block. According to the block type, bilinear interpolation (BI), directional interpolation (DI) or best neighborhood matching (BNM) is used for the error concealment of the block. At encoder end, we apply disperse mode of FMO to ensure that the four-neighbors of the lost macroblock are present and hide block type message into its neighbor block. At the decoder, we extract the hided data and utilize them to decide the block should apply the appropriate error concealment algorithm. The experiment result shows that proposed method has better subjective perception than earlier works.

[1] D. Agrafiotis, D. R. Bull, C. N. Canagarajah, “Enhanced Error Concealment With Mode Selection,” IEEE Trans. on Circuits and Systems for Video Technology , vol. 16, no. 8, pp. 960 – 973, 2006.
[2] X. Xiu, L. Zhuo, L. Shen, “A hybrid error concealment method based on H.264 standard,” The 8th International Conference on Signal Processing, vol. 2, pp. 16-20, 2006.
[3] M. J. Chen, W. W. Liao, M. C. Chi, ”Robust temporal error concealment for H.264 video decoder,” Digest of Technical Papers. International Conference on Consumer Electronics, pp. 383 – 384, Jan. 2006.
[4] K. Donghyung, Y. Siyoung, J. Jechang, “A new temporal error concealment method for H.264 using adaptive block sizes”, IEEE International Conference on Image Processing, vol. 3, no.3, pp. 928-931, Sept. 2005.
[5] E. El Khamy, M. A. Qabel, “A New Hybrid Classification-Based Adaptive Error Concealment Technique for Image Transmission through Wireless Communication Channels,” National Radio Science Conference,pp. 1-9, 13-15 March 2007.
[6] Yan Chen, Oscar Au, Jiantao Zhou, Chi-Wang Ho, “Adaptively Switching Between Directional Interpolation and Region Matching for Spatial Error Concealment Based on DCT Coefficients,” IEEE International Conference on Multimedia and Expo, Page(s):805 – 808, July 2006.
[7] Zhang Rongfu, Zhou Yuanhua and Huang Xiaodong, “Content-adaptive spatial error concealment for video communication,” , IEEE Transactions on Consumer Electronics, Volume 50, Issue 1, Page(s):335 – 341, Feb. 2004.
[8] L.-W. Kang and J.-J. Leou, “A hybrid error concealment scheme for MPEG-2 video transmission based on best neighborhood matchingalgorithm,” J. Vis. Commun. Image Represent., vol. 16, no. 3, pp. 288–310, Jun. 2005.