Fast Extraction of Edge Histogram in DCT Domain based on MPEG7

In these days, multimedia data is transmitted and processed in compressed format. Due to the decoding procedure and filtering for edge detection, the feature extraction process of MPEG-7 Edge Histogram Descriptor is time-consuming as well as computationally expensive. To improve efficiency of compressed image retrieval, we propose a new edge histogram generation algorithm in DCT domain in this paper. Using the edge information provided by only two AC coefficients of DCT coefficients, we can get edge directions and strengths directly in DCT domain. The experimental results demonstrate that our system has good performance in terms of retrieval efficiency and effectiveness.


Keywords:


References:
[1] K. R. Rao, J. J. Hwang, " Techniques and Standards for image, Video, and
Audio Coding, " Prentice Hall PTR, 1996.
[2] Overview of the MPEG-7 standard, December 2001. ISO/IEC/TC/SC29/
WG11 N3914
[3] B.S. Manjunath, Philippe Salembier, Thomas Sikora, "Introduction to
MPEG-7", JOHN WILLEY & SONS, LTD , pp. 183-184, 2002
[4] Bo Shen, Ishwar K. Sethi, ÔÇÿDirect feature extraction from compressed
images-, SPIE vol. 2670, Storage & Retrieval for Image and Video
Databases , 1996.
[5] S. W. Lee, Y.M. Kim and S. W. Choi, ÔÇÿFast scene change detection using
direct feature extraction form MPEG compressed videos-, IEEE Trans.
On Multimedia, Vol. 2, No. 4, December 2000