視訊內容識別之適應影像傳輸

Content aware adaptive transmission for streaming video

蔡秉樸(Bing-Pu Tsai)

指導教授 ： 陳永昌

國立清華大學/電機資訊學院/電機工程學系/碩士(2006年)

摘要

最近以來，在無線網路上傳輸的影像品質一直是許多影像研究人員努力加強的重點，由於在網路上的即時視訊應用增多，多媒體使用者的需求也越來越高，然而影像封包在網路中傳輸時，常會遇到網路壅塞或是頻寬不穩的問題，往往造成封包的遺失，它不僅會影響該畫面的品質，還會造成錯誤傳遞到之後的畫面，這是因為視訊編碼方式採用了運動補償技術的緣故。
所以封包遺失是一個棘手的問題，為了解決此問題，錯誤追蹤 (error tracking) 數年前已經有人提出，它主要的想法就是先將估計畫面錯誤蔓延的情況，然後將嚴重錯誤的區塊更新編碼，如果封包在傳輸時，萬一遺失而造成畫面受損，intra編碼區塊即可有效的中止錯誤蔓延，並且接續畫面品質仍可以快速的回復。在這篇論文中我們提出一個新的演算法，將畫面上的錯誤定位並且快速的估計錯誤的蔓延，使得區塊更新編碼更有效率。這裡提出的演算法可以因應影像內容重要性和遺失封包，來調控其更新編碼率，為了即時傳輸上的應用，錯誤追蹤的速度也必須考慮進去。
在本論文中，我們的模擬結果展示在不同的錯誤蔓延長度下的錯誤追蹤和追蹤可調性的表現，我們設定的intra更新門檻可以因應錯誤區塊，使得intra更新編碼做得更有效率，在高準確度的錯誤估計下，錯誤追蹤的可調性依然能夠做到不錯的水準。

關鍵字

適應傳輸

並列摘要

Recently, many people keep working with an emphasis on enhancement of the video quality on wireless network. The real-time video application on network is increasing and the demand of users growing. However, when the video packets are transmitted on network, they may be lost due to network congestion or unstable bandwidth. It not only affects the quality of current frame, but also leads to error propagation to subsequent frames because of motion compensated prediction techniques used in standard video codecs.
The loss of packets is a serious problem. To solve it, error tracking was proposed several years ago, the main idea is to predict the propagation of errors and encode some blocks as intra block. If we lost some packets when transmitting, the intra blocks will stop the error propagation and recover the image quality quickly. Here, we propose a new algorithm to locate the error and tracking the error propagation quickly. It enhances the efficiency of the intra refresh coding. The proposed algorithm can adaptively vary the intra refresh rate according to the video content and the packet loss to protect the most important macroblocks against data loss. For real-time application, the speed of error tacking is considered.
In this thesis, we show the simulation results of the error tracking algorithms with different propagation length and the performance of tracking scalability. The setting of intra-refresh threshold is adaptive to erroneous MBs and the intra-refresh becomes more effective. Due to the high accuracy of error prediction, the scalability of error tracking still works well.

並列關鍵字

adaptive transmission

參考文獻

[1] W.-H. J. Chen, J.-N. Hwang, The CBERC: A content-based error-resilient coding technique for packet video communications, IEEE Trans. Circuits Systems Video Technol. 11 (8) (2001) 974-980.

[2] Abdul H. Sadka, Compressed Video Communications, 2002.

[3] H.-J. Chiou, Y.-R. Lee, and C.-W. Lin, “Error-resilient transcoding using adaptive intra-refresh for video streaming,” in Proc. IEEE Int. Symp. Circuits and Systems, vol. 3, pp. 777-780, May 2004, Vancouver, Canada.

[4] H.-J. Chiou, Y.-R. Lee, and C.-W. Lin, “Content-aware error resilient transcoding using prioritized intra-refresh for video streaming,” Journal of Visual Communication and Image Representation, vol. 16, no. 4-5, pp. 563-588, Aug.-Oct. 2005.

[5] E. Steinbach, N. Färber, B. Girod, Standard compatible extension of H.263 for robust video transmission in mobile environments, IEEE Trans. Circuits Syst. Video Technol., vol. 7, pp. 872-881, Dec. 1997.