跳到主要內容

臺灣博碩士論文加值系統

(54.225.48.56) 您好!臺灣時間:2022/01/19 21:08
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

: 
twitterline
研究生:楊行健
研究生(外文):Hsing-Chien Yang
論文名稱:MPEG-2視訊編碼之場景轉換偵測
論文名稱(外文):Scene-Change Detection for MPEG-2 Video Encoder
指導教授:戴顯權戴顯權引用關係
指導教授(外文):S.C.Tai
學位類別:碩士
校院名稱:國立成功大學
系所名稱:電機工程學系碩博士班
學門:工程學門
學類:電資工程學類
論文種類:學術論文
論文出版年:2002
畢業學年度:90
語文別:英文
論文頁數:58
中文關鍵詞:視訊編碼小波轉換場景偵測
外文關鍵詞:MPEGWaveletsScene-Change
相關次數:
  • 被引用被引用:0
  • 點閱點閱:271
  • 評分評分:
  • 下載下載:35
  • 收藏至我的研究室書目清單書目收藏:1
MPEG-2視訊編碼之場景轉換偵測

楊行健* , 戴顯權**
國立成功大學電機研究所

摘要
近年來,視訊壓縮的技術在許多的應用上扮演著相當重要的角色,例如資料傳輸與資料儲存。MPEG-2標準是為了能提供更好的品質與更具彈性的應用而制定的。雖然MPEG-2視訊編碼標準能提供高品質的視覺效果,但如果有場景轉換發生時整體的編碼效能與視覺效果仍會下降。為了解決這個問題我們視訊編碼器提出一種新的場景轉換偵測機制以避免畫質的低落。我們提出一個以運動補償為基礎的演算法用來量測兩張相鄰畫面的相似性。為了更減低計算量我們使用小波轉換後最低頻的信號去判斷是否有場景轉換。我們的演算法可以避免兩張完全不同畫面卻具有相同特徵時所造成的誤判。實驗結果顯示所提出演算法就畫面品質與錯誤判斷上都具有相當好的效果。我們也發現所提出的演算法和傳統MPEG-2標準在相同的位元速率下畫質大約可以有0.33 dB的提昇。因此這篇論文所提出的演算法可以應用於視訊資料壓縮與視訊資料庫索引等方面。

*作者 **指導教授
Scene-Change Detection for MPEG-2 Video Encoder

Hsing-Chien Yang *, Shen-Chuan Tai**
Department of Electrical Engineering
National Cheng Kung University, Tainan, Taiwan, R.O.C

Abstract
In recent years, video compression technique plays an important role in many applications such as data transmission and storage. MPEG-2 standard is instituted for better visual quality and more flexible applications. Although the MPEG-2 video standard can provide high quality, the performance will decrease if a scene-change occurs. To solve this problem, we propose a novel scene-change detection scheme for video encoder to prevent visual quality decrease in uncompressed domain. We present a motion-compensation based algorithm to measure the similarity of two adjacent frames. To further reduce the computational load, we use the lowest band of wavelet transform to estimate the occurrence of scene-change. Our algorithm can present false detection when two different images have the same feature. Experimental results show that the performance of the proposed algorithm is indeed superb in term of visual quality and false alarm. Also, we found that the proposed algorithm has the better visual quality than MPEG-2 standard about 0.33 dB at the same bitrate. As a result, this contribution could be applied in the video applications such as database indexing or data compression.

*Author **Advisor
CONTENTS

List of Figures
List of Tables

Chapter 1 Introduction 1
Chapter 2 MPEG-2 Video Coding 3
2.1 I-picture Coding 4
2.2 P-picture Coding 5
2.3 B-picture Coding 8
2.4 Scalable Coding Techniques in MPEG-2 9
2.4.1 Spatial Scalability 10
2.4.2 SNR Scalability 12
2.4.3 Temporal Scalability 13
2.4.4 Data Partitioning 14
Chapter 3 Scene-Changes Detection 15
3.1 Types of Scene-Change 15
3.2 Abrupt Scene-Change Detection Algorithms 17
3.2.1 Pixel-Based Methods 18
3.2.2 Histogram-Based Methods 19
3.2.3 Edge-Based Methods 20
Chapter 4 Proposed Algorithm 22
4.1 Influence of Scene-Changes 22
4.2 Our Proposed Method 23
Chapter 5 Experimental Results and Comparison 31
5.1 Filter Banks and Motion Estimation 31
5.2 Video Sequence and Experimental Environment 32
5.3 Experimental Results 32
Chapter 6 Conclusions and Future Work 45
Reference 59




List of Figures

Page
Figure 2-1. A typical GOP structure…………………………………………………..4
Figure 2-2. Function block diagram for I-picture coding type………………………..4
Figure 2-3. Function block diagram for P-picture coding type……………………….6
Figure 2-4. Frame-prediction for P-pictures.………………………………………….6
Figure 2-5. Field prediction for P-picture. (a) Prediction of the first filed for P-pictures; (b) Prediction of the second filed for P-pictures when it is the bottom field. (c) Prediction of the second filed for P-pictures when it is the top field………………………………………………………………………...9
Figure 2-6. Function block diagram for B-picture coding type…………………….…8
Figure 2-7. Field-prediction of B field pictures…………………………………….…9
Figure 2-8. Frame-prediction for B-pictures………………………………………….9
Figure 2-9. Function block diagram for video encoder with spatial scalability……...11
Figure 2-10. Function block diagram for video encoder with SNR scalability……...13
Figure 2-11. A temporal scalability structure………………………………………...14
Figure 2-12. An example of data partitioning………………………………………..14
Figure 3-1. Table Tennis 66th frame…………………………………………………..15
Figure 3-2. Table Tennis 67th frame…………………………………………………..15
Figure 3-3. A fad-out example………………………………………………………..16
Figure 3-4. A fad-in example…………………………………………………………16
Figure 3-5. An example of dissolving………………………………………………..17
Figure 3-6. An example of wiping…………………………………………………...17
Figure 3-7. Hierarchical diagram of abrupt scene-change detection…………………18
Figure 4-1. The influence of a scene change: (a) I-frame is a scene-change frame, (b) P-frame is a scene-change frame, (c) B-frame is a scene-change frame, and (d) B-frame is a scene-change frame………………………………24
Figure 4-2. Phase 1: Inter-Frame Similarity Measurement…………………………..25
Figure 4-3. An example of inter-frame similarity measurement. (a),(b),(c) and (d) are input sequences. (e) motion-compensated by (a) and (b). (f) motion-compensated by (b) and (c). (g) motion-compensated by (c) and (d). (g) PSNR vs frame No……………………………………………….29
Figure 4-4. Illustration of a sliding window approach……………………………….29
Figure 4-5. Four different pictures with the same histogram distribution……………30
Figure 5-1. Spatial wavelet decomposition scheme resulting in 7 subbands………...31
Figure 5-2. Inter-frame similarity measurement of test video sequence 1…………...35
Figure 5-3. PSNR performance of test sequence 1 at 5Mbits/s and ±15.5 motion estimation search range both P and B frames…………………………….35
Figure 5-4. PSNR performance of test video sequence 1 at 5Mbits/s and ±63.5 motion estimation search range both P and B frames……………………36
Figure 5-5. Inter-frame similarity measurement of test video sequence 2…………...36
Figure 5-6. PSNR performance of test video sequence 2 at 5Mbits/s and ±15.5 motion estimation search range both P and B frames……………………37
Figure 5-7. PSNR performance of test sequence 2 at 5Mbits/s and ±63.5 motion estimation search range both P and B frames…………………………….37
Figure 5-8. Inter-frame similarity measurement of test video sequence 3…………...38
Figure 5-9. PSNR performance of test video sequence 3 at 5Mbits/s and ±15.5 motion estimation search range both P and B frames……………………38
Figure 5-10. Subband-DOH for test video sequence 1………………………………40
Figure 5-11. Subband-CHI for test video sequence 1………………………………..40
Figure 5-12. CHI for test video sequence 1………………………………………….41
Figure 5-13. Subband-DOH for test video sequence 2………………………………41
Figure 5-14. Subband-CHI for test video sequence 2………………………………..42
Figure 5-15. CHI for test video sequence 2…………………………………………..42
Figure 5-16. Subband-DOH for test video sequence 3………………………………43
Figure 5-17. Subband-CHI for test video sequence 3………………………………..43
Figure 5-18. CHI for test video sequence 3…………………………………………..44


List of Tables

Page
Table 5-1 Composition of test sequence 1………………………………………….33
Table 5-2 Composition of test sequence 2………………………………………….34
Table 5-3 Composition of test sequence 3………………………………………….34
Table 5-4. The performance’s comparison between with and without Scene-Change Detection for test video sequence 1……………………………………..39
Table 5-5. The performance’s comparison between with and without Scene-Change Detection for test video sequence 2……………………………………..39
Table 5-6. The performance’s comparison between with and without Scene-Change Detection for test video sequence 3……………………………………..39
REFERENCE

[1] J. Lee and B. W. Dickinson,“Temporally adaptive motion interpolationexploiting temporal masking in visual perception,”IEEE Trans. Image Processing, vol. 3, pp. 513-526, Sept. 1994.
[2] J. Lee and B. W. Dickinson,“Hierarchical Video Indexing and Retrieval for Subband-Coded Video,” IEEE Trans. Circuit Syst. Video Technol., vol. 10, No. 5, pp. 824-829, Aug. 2000.
[3] Li-Jun Luo, Cai-Rong Zou, and Zhen-Ya He, “A New Algorithm on MPEG-2 Target Bit-Number Allocation at Scene Changes,” IEEE Trans. Circuit Syst. Video Technol., vol. 7, No. 5, pp. 815-819, Oct. 1997.
[4] Kuo-Chin Fan and Kuo-Sou Kan,“An Active Scene Analysis-Based Approach for Pseudoconstant Bit-Rate Video Coding” IEEE Trans. Circuit Syst. Video Technol., vol. 10, No. 7, pp. 159-1184, April 1998.
[5] (1996) MPEG Software Simulation Group, Test model simulator, version 1.2. [Online]. Availabel: http://www.mpeg.org/MPEG/
[6] Jeff McVeigh, George K. Chen, Judi Goldstein, Atul Gupta, Mike Keith, and Steve Wood, “A Soft-Based Real-Time MPEG-2 Encoder”IEEE Trans. Circuit Syst. Video Technol., vol. 10, No. 7, pp. 1178-1184, Oct. 2000.
[7] Ullas Gargi, Rangachar Kasturi, and Susan H.Strayer“Performance Characterization of Video-Shot-Change Detection Methods,” IEEE Trans. Circuit Syst. Video Technol., vol. 10, No. 1, pp. 1-13, Feb. 2000.
[8] H. J. Zhang, A.Kankanhalli, and S. W. Smoliar, “Automatic partitioning of full-motion video,”Multimedia System, vol. 1, pp 10-28, July 1993.
[9] Byung Cheol Song, Myung Jun Kim, and Jon Beom Ra,“A Fast Multiresolution Feature Matching Matching Algorithm for Exhaustive Search in Large Image Database,”IEEE Trans. Circuit Syst. Video Technol., vol. 11, No. 5, pp. 673-678, May. 1997.
[10] H. J. Zhang, Gio Wiederhold, Oscar Firschein, and Sha Xin Wei,“Wavelet-Based Image Indexing Techniques with Partial Sketch Retrieval Capability,”IEEE Proc. Fourth Forum on Research and Technology Advances in Digital Libraries, 1997.
[11] Haitao Jiang, Abdelsalam (Sumi) Helal, Ahmed K. Elmagarmid, Anupam Joshi “Scene change diction techniques for video database system,” Multimedia System, vol. 6, pp 186-195, July 1998.
[12] X. Wen, T. D. Huffmire, H. H. Hu and A. Finkelestein,“Wavelet-based video indexing and querying,”Multimedia Systems, vol. 7 1999, pp. 350-358.
[13] A. Nagasaka and Y. Tanaka, “Automatic video indexing and full video search for object appearances,”in Proc. IFIP 2nd Working Conf. Visual Database System, 1992, pp. 113-127.
[14] MPEG-2 Video Test Model 5, ISO/IEC JTC1/SC29/WG11, Doc. N0400, Apr 1993.
[15] Information Technology-Generic Coding of Moving Picture and Associated Audio Information: Video, IOS/IEC 13818-2:1996(E), 1996.
[16] B. L. Yeo and B. Liu, “ Rapid scene analysis on compressed video,”IEEE Trans. Circuit Syst. Video Technol., vol. 5, pp. 533-544, 1995.
[17] Ulrich Benzler, “ Spatial Scalable Video Coding Using a Combined Subband-DCT Approach,”IEEE Trans. Circuit Syst. Video Technol., vol. 10, No. 7, pp. 1080-1087, Oct. 2000.
[18] J. Lee and B. W. Dickinson,“Subband Video Coding with Scene-Adaptive Hierarchical Motion Estimation,” IEEE Trans. Circuit Syst. Video Technol., vol. 9, No. 3, pp. 459-466, April. 2000.
[19] Chung-Lin Huang and Bing-Yao Liao,“A Robust Scene-Change Detection Method for Video Segmentation,” IEEE Trans. Circuit Syst. Video Technol., vol. 11, No. 12, pp. 1281-1288, Dec. 2001.
[20] Ramin Zabih, Justin Miller, Kevin Mai, “A feature-based algorithm for detecting
and classifying production effects,” Multimedia Systems, vol. 7 1999, pp. 119-128.
[21] J. Lee and B. W. Dickinson, ``Rate-distortion optimized frame type selection for MPEG encoding,', IEEE Trans. Circuit Syst. Video Technol. 7, 1997, pp. 501-510.
連結至畢業學校之論文網頁點我開啟連結
註: 此連結為研究生畢業學校所提供,不一定有電子全文可供下載,若連結有誤,請點選上方之〝勘誤回報〞功能,我們會盡快修正,謝謝!
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top