臺灣博碩士論文加值系統

新的視訊標準「MPEG-4」，提供了以物件為基礎的壓縮方式，在以物件為基礎的編碼當中，物件的形狀 (arbitrary shape) 與物件的內容(texture) 將被編碼於資料串流 (bitstream) 之中。當物件的形狀資訊在傳輸中被破壞時，會造成整個物件在解碼時的錯誤與不正確。因此，在這篇論文裡我們提出了當錯誤發生時，如何去找出錯誤的位置並發展錯誤隱藏的演算法。
在尋找錯誤發生的方法上，MPEG-4已經提供了很多的輔助錯誤復原的工具(error resilience tools), 利用這些工具與一些在形狀資料編碼上的特性，我們可以偵測大部份錯誤發生的位置，當錯誤的位置被找到之後，可用錯誤隱藏技術來改善影像品質，在視訊的空間域 (spatial domain) 部份，我們利用被破壞區域周圍鄰近的點與其對應的切線斜率，提出了貝茲曲線趨近法 (Bezier Curve Fitting，BCF) ；而在視訊的時間域 (temporal domain) 部份，我們利用被破壞區域周圍鄰近的點並給與其加權值而提出了加權邊際比對 (Weighted Side Matching，WSM) 準則，去由前一張的畫面找出一個最適當區塊來隱藏錯誤。而這兩個方法都在物件形狀的邊界上能得到不錯平滑的曲線與連續無斷的接合點，在客觀與主觀上都有很好的品質。

The new generation of video coding standard – MPEG-4 provides applications for mobile applications. The MPEG-4 international standard supports the object-based coding, which provides the capability to directly access arbitrarily shaped video object. The shape information is also coded in data streams. When the compressed data streams are transmitted, data packets may be lost or corrupted, due to either traffic congestion or bit errors caused by impairment of the physical channels. The transmission error will cause distortions in decoded pictures.
In this thesis, new methods for error concealment of shape information are proposed for compressed MPEG-4 bit-streams transmitted over error prone channels. The error resilient tools in MPEG-4 are used to detect the errors. When errors have been detected, we use our proposed algorithms to conceal these errors. For spatial domain, the proposed approach employs the Bezier curves fitting to approximate and outline the arbitrary shape. The Bezier curve is the representation that is the most frequently used in computer graphics. In general, a Bezier curve section can be fitted to any number of control points. Using the Bezier curve feature, we modify and simplify the formula of Bezier curve for the application of error concealment. For temporal domain, the proposed approach uses the weighted side matching criterion to reconstruct VOP. The proposed criterion can find the best matched block in previous frame. The best matched block is used to reconstruct the damaged block. They get better subjective and objective qualities than previous works.

Abstract 1
Chapter 1 Introduction 3
1.1 Concept of MPEG-4 Video Coding 3
1.2 Shape Coding 6
1.3 Organization of the Thesis 9
Chapter 2 Error Detection of MPEG-4 Shape Coding 10
2.1 Overview of Error Resilience Coding Tools 10
2.2 Proposed Error Detection Algorithm 12
2.2.1 The Feature of Errors in Bitstream 12
2.2.2 Bitstream Domain 15
2.2.3 Spatial Domain 19
Chapter 3 Spatial Error Concealment Algorithm 21
3.1 Review of Error Concealment Methods in Spatial Domain 22
3.1.1 Maximum a Posteriori (MAP) Estimator 22
3.1.2 Fuzzy-based Method 24
3.2 Proposed Spatial Error Concealment Algorithm 27
3.2.1 Error Block Types 27
3.2.2 Labeling The Error Blocks 28
3.2.3 The Control Points P0, P1 and P2 29
3.2.4 Bezier Curve Theory 31
3.2.5 Bezier Curve Fitting (BCF) Algorithm 38
3.2.6 Final Step 42
3.3 Experimental Results 42
Chapter 4 Temporal Error Concealment Algorithm 54
4.1 Review of Error Concealment Methods in Temporal Domain 57
4.1.1 Smooth Function Candidate (SFC) 57
4.1.2 Decoder Motion Vector Estimation (DMVE) 59
4.2 Proposed Temporal Error Concealment 61
4.2.1 Macroblock vs. Binary Alpha Block 61
4.2.2 Weighted SAD 63
4.2.3 Weighted Side Matching 67
Chapter 5 Adaptive Error Concealment 84
Chapter 6 Conclusion 90
Bibliography 93

[1] Coding of moving pictures and audio: Video, ISO/ICE JTC1/SC29/WG11, March 2000.

[2] M. Grube, P. Siepen, C. Mittendorf, M. Boltz, M. Srinivasan, “Applications of MPEG-4: digital multimedia broadcasting,“ IEEE Transactions on Consumer Electronics, Vol.47, pp. 474–484, Aug. 2001.

[3] A.Vetro, Huifang Sun, ” An overview of MPEG-4 object-based encoding algorithm,” International Conference on Information Technology: Coding and Computing, pp. 366–369, 2001.

[4] K.A. Jacob, J. Hormigo, ” Bit stream processor for object based MPEG-4 profiles,” Conference Record of the Thirty-Fifth Asilomar Conference on Signals, Systems and Computers, pp. 1241–1245, 2001.

[5] Hao-Chieh Chang, Yi-Chu Wang, Mei-Yun Hsn, Liang-Gee Chen, ”Efficient algorithms and architectures for MPEG-4 object-based video coding,” IEEE Workshop on Signal Processing Systems, pp. 13–22, 2000.

[6] R.M Guirguis, S. Mahmoud, “An intelligent VOP discard algorithm(IVDA) for transmission of multi-layered MPEG-4 video over explicit rate networks,” IEEE Global Telecommunications Conference, pp. 2494–2500, 2001.

[7] N. Brady, F. Bossen, N. Murphy, “Context-based arithmetic encoding of 2D shape sequences,” International Conference on Image Processing, Vol. 1, pp. 29–32, 1997.

[8] N. Brady, F. Bossen, N. Murphy, “Context-based arithmetic encoding of 2D shape sequences,” International Conference on Image Processing, Vol. 1, pp.29-32, 1997.

[9] J. Ostermann, “Efficient encoding of binary shapes using MPEG-4,” International Conference on Image Processing, Vol. 1, pp. 295-298, 1998.

[10] Wang Yao, S. Wenger, Wen Jiantao, A.K. Katsaggelos, “Error resilient video coding techniques,” IEEE Signal Processing Magazine, Vol.17, pp.61-82, July 2000.

[11] S. Shirani, B. Erol, and F. Kossentini, “A concealment method for shape information in MPEG-4 coded video sequences,” IEEE Transactions on Multimedia, Vol.2, No.3, pp. 185-190, September 2000.

[12] Pei-Jun Lee, Liang-Gee Chen, Wen-June Wang, and Mei-Juan Chen, “Robust error concealment algorithm for MPEG-4 with the aids of fuzzy theory,” IEEE International Conference on Consumer Electronics, Los Angeles, U.S.A., pp. 154-155, June 2001.

[13] M.R. Frater, W.S Lee, M. Pickering and J.F. Arnold, “Error concealment in video coding of arbitrarily shaped objects,” Signal Processing: Image Communication, Vol. 15, pp. 631-641, 2000.

[14] Chien-Ming Lee, “Detection and concealment of transmission errors in MPEG-4 images,” Workshop on Consumer Electronic, Taipei, Taiwan, pp.3A061-3A067, Oct. 2001.

[15] J.L.H. Webb, “Efficient table access for reversible variable-length decoding,” IEEE Transactions on Circuits and Systems for Video Technology, Vol. 11, pp. 981-985, Aug. 2001.

[16] Yue-Lin Lee, Chung-Neng Wang, and Tihao Chiang, “An MPEG-4 error resilient decoder,” Workshop on Consumer Electronic, Taipei, Taiwan, pp.5B241-5B245, 2001.

[17] Donald Hearn and M. Pauline Baker, Computer Graphics, Prentice Hall, pp 327-331.

[18] S. Geman and D. Geman, “Stochastic relaxation, Gibbs distribution, and the Bayesian restoration of images,” IEEE Transactions on Pattern Anal. Machine Intell., Vol. PAMI-6, pp.721-741, Nov. 1984.

[19] P.J. Lee and M.J. Chen, ”Robust error concealment algorithm for video decoder,” IEEE Transactions on Consumer Electronic, Vol. 45, No. 3, pp.851-859, August 1999.

[20] M.J. Chen, L.G. Chen, and R.M. Weng, “Error concealment of lost motion estimation vectors with overlapped motion compensation,” IEEE Transactions on Circuits and Systems for Video Technology, Vol. 7, No. 3, pp.560-563, June1997.

[21] W.M. Lam, Reibman A.R., and Liu B., “Recovery of lost or erroneously received motion vectors,” Proc. ICASSP, Vol. 5, pp. 417-420, 1993.

[22] M. Ghanbari, “Cell-loss concealment in ATM video codes,” IEEE Transactions on Circuits and Systems for Video Technology, Vol. 3, pp. 238-247, June 1993.

[23] Chien-Ming Lee, “Detection and concealment of transmission errors in MPEG-4 images, ” Workshop on Consumer Electronic, Taipei, Taiwan, pp.3A061-3A067, Oct. 2001.

[24] Jian Zhang, J.F. Arnold, M.R. Frater, M.R. Pickering, “Video error concealment using decoder motion vector estimation,” IEEE Region 10 Annual Conference on Speech and Image Technologies for Computing and Telecommunications., Vol. 2, pp. 777-780, 1997.