Dissertation > Excellent graduate degree dissertation topics show

Research on Video Shot Boundary Detection in H.264/AVC Compressed Domain

Author: XieHuiLin
Tutor: XiaDingYuan
School: Wuhan University of Technology
Course: Communication and Information System
Keywords: H.264/AVC standard shot boundary detection mode selection prediction mode threshold
CLC: TN919.8
Type: Master's thesis
Year: 2009
Downloads: 36
Quote: 0
Read: Download Dissertation

Abstract


Content-based video retrieval technology aims at making summary, browse and retrieval of the digital video by using shot boundary detection, key-frame extraction, feature extraction, feature matching and other methods, in order to manage the video information effectively. As the key technology of content-based video retrieval system, shot boundary detection attracts many researchers’ attention. The traditional shot boundary detection was applied to the pixel domain, using the difference information of the pixel between frames. And the methods based on MPEG compressed domain can segment shot boundary effectively by adopting the characteristic of the compressed video.As the new generation video coding standard, H.264/AVC becomes very popular for its efficient performance in video compression and network applicability. Since there is a lot of new technologies adopted by H.264/AVC, such as intra-prediction, multi-size block segmentation, multiple reference frames and so on, the former shot boundary detection methods can not be directly applied to the H.264/AVC compressed stream. For this reason, this paper studies the shot boundary detection technology in H.264/AVC compressed domain.Firstly, this paper analyzed the classic shot boundary detection algorithm based on pixel domain and MPEG compressed domain. Then, it discussed the development of shot segmentation based on H.264/AVC compressed domain. Finally, the paper presented a new shot boundary detection algorithm based on H.264/AVC compressed domain named PMD (A Shot Boundary Detection Algorithm Based on Prediction Mode Difference).The prediction mode information of frames is extracted, and the difference of corresponding blocks mode between frames is calculate. At last, the various types of shot boundary are detected by setting adaptive threshold. The main innovation is mainly reflected in the following:(1)Aiming at some defect in pixel domain and the MPEG compressed domain for shot boundary detection, making full use of new characteristics of the H.264/AVC video coding standard, a new shot boundary detection method based on H.264/AVC compressed domain using difference information of prediction mode between frames is proposed. At first, the block size in frames should be normalized, and then the difference between the corresponding blocks in the intra-prediction mode or the inter-prediction is counted up to detect the shot boundary. The new method brings more effective detection but reduces the computational complexity.(2)Aiming at the difficulty of shot boundary detection in all types of coded frames, a unified detection model is established. It can not only detect the shot boundary in I frame, but also can point out the precise position of the shot transition.The experimental results show that the proposed algorithm achieves satisfactory performance of recall and precision in both abrupt transition and gradual transition.

Related Dissertations

  1. Resrarch on Multisignatures and Multisigncryptions from Identity-based,TN918.1
  2. Research on Spread Spectrum Code Acquisition of Double Threshold Based on Sliding Correlator,TN914.42
  3. Performance Analysis of Anti-Multipath Capability in DSSS System and Study on the Compensation Methods,TN914.42
  4. Citrus Image Segmentation Based on Genetic Algorithm,TP391.41
  5. Macroinvertebrate Assemblages and Its Application in the Water Management,X824
  6. Bioavailability of Phthalic Acid Esters in Different Types of Soil-plant Systems,X53
  7. Differences in different populations compared proprioception and Brain Mechanisms,B845
  8. Subliminal subliminal anxiety of material of different emotional cues students Inhibition of Return,B842
  9. Mobile WSN data collection based on the virtual cluster head Strategy,TP212.9
  10. Research on Face Detection Based on Skin Color Segmentation and AdaBoost Algorithm,TP391.41
  11. Study on Skyline Query Processing Techniques on Uncertain Data,TP311.13
  12. The Study of Human Face Detection Method Based on OpenCV,TP391.41
  13. Design and Implementation of Wireless Mesh Network Testbed Based on Network Cards,TN929.5
  14. Research of Scalable Video Coding Based on H.264/AVC Extension,TN919.81
  15. Research on Mode Decision Algorithm for H.264/AVC Intra Prediction,TN919.81
  16. Research on PAPR Technology of Wavelet Packet Multicarrier Modulation System,TN919.3
  17. Research on Methods of Image Processing of the Image Information Processor,TP391.41
  18. Research of Facial Expression Recognition Algorithm,TP391.41
  19. Cloud Classification Based on Geostationary Meteorological Satellite Imagery,TP391.41
  20. Research for Sbs Threshold Enhanced in Single-mode Fiber,O437.2
  21. Study on Nonwoven Fabric Defects Inspection,TP391.41

CLC: > Industrial Technology > Radio electronics, telecommunications technology > Communicate > Image communication, multimedia communication
© 2012 www.DissertationTopic.Net  Mobile