Comparative Study on Recent Integer DCTs

This paper presents comparative study on recent integer DCTs and a new method to construct a low sensitive structure of integer DCT for colored input signals. The method refers to sensitivity of multiplier coefficients to finite word length as an indicator of how word length truncation effects on quality of output signal. The sensitivity is also theoretically evaluated as a function of auto-correlation and covariance matrix of input signal. The structure of integer DCT algorithm is optimized by combination of lower sensitive lifting structure types of IRT. It is evaluated by the sensitivity of multiplier coefficients to finite word length expression in a function of covariance matrix of input signal. Effectiveness of the optimum combination of IRT in integer DCT algorithm is confirmed by quality improvement comparing with existing case. As a result, the optimum combination of IRT in each integer DCT algorithm evidently improves output signal quality and it is still compatible with the existing one.

Authors:

• Sakol Udomsiri
• Masahiro Iwahashi

Keywords:

• DCT
• sensitivity
• lossless
• wordlength.

References:

[1] W. B. Pennebaker and J. L. Mitchell, JPEG still image data compression
standard, Van Nostrand Reinhold, NY, 1993.
[2] J. L. Mitchell, W.B. Pennebaker, C. E. Fogg, and D. J. LeGall, MPEG
video compression standard, Chapman and Hall, NY, 1997.
[3] K. R. Rao and J.J. Hwang, Technique and standards for image, video
and audio coding, NJ: Prentice Hall Inc., 1996.
[4] Fukuma S., Ohyama K., Iwahashi M., Kambayashi N., Lossless
8-Point Fast Discrete Cosine Transform Using Lossless Hadamard
Transform, Technical report of IEICE, DSP99-103 (Oct.1999) 37-44.
[5] S. CHOKCHAITAM, M. IWAHASHI and S. JITAPUNKUL, A New
Unified Lossless/Lossy Image Compression based on A New Integer
DCT, IEICE Trans. On Information system (July 2005) 465-478.
[6] K. Komatsu, K. Sezaki, 2D Lossless Discrete Cosine Transform, (IEEE
ICIP 2001)466-469.
[7] J. Liang and T. D. Tran, Fast multiplierless approximations of the DCT
with the lifting scheme, IEEE Transactions on Signal Processing, Vol.
49, No.12, December 2001, pp. 3032-3044.
[8] Y. J. Chen, S. Oranintara and T. Nguyen, Integer discrete cosine
transform (IntDCT), Proceedings of the 2nd Internationals Conference
on Information, Communications and Signal Processing, Singapore,
December 1999.
[9] M.D. Adams, F. Kossentini, Reversible Integer-to-Integer Wavelet
Transform for image Compression: Performance Evaluation and
Analysis, IEEE Transactions on Image Processing, vol.9, no. 6 (June
2000) 1010-124.
[10] S. Chockchaitam, and M. Iwahashi, Performance Evaluation of the
Lossless/Lossy Wavelet for Image Compression under Lossless/Lossy
Coding Gain, IEICE special section on "Digital Signal Processing", vol.
E85-A, no. 8 (Aug. 2002) 1882-1891.
[11] M.D.Adams, F. Kossentini, Reversible integer-to-integer wavelet
transform for image compression: Performance evaluation and analysis,
IEEE Trans, Image Process. 9 (6) (June 2000)1010-1024.
[12] S. Chokchaitam, M. Iwahashi and N. Kambayashi, Optimum word
length allocation of integer DCT and its error analysis, J.Sci. Signal
Processing: Image Communication, 19( 2004) 465-478.
[13] M. Iwahashi, O.Nishida, S. Chockchaitam, N.Kambayashi, Optimum
Word Length Allocation for Multipliers of Integer DCT, IEEE ISCAS
2003, vol.11 (May 2003) 400-403.
[14] D.K.Dang, M.Ohnishi, M.Iwahashi, S. Chochaitam, A New Structure
of Integer DCT Least Sensitive to Finite Word Length Expression of
Multipliers, IEEE International Conference on Image Processing
(ICIP), No.II, (Sep.2005) 269-272.
[15] V. Britanak, P. C. Yip and K. R. Rao, Discrete cosine and sine
transforms, first ed. Academic Press is an imprint of Elsevier, UK,
2007.
[16] G. Plonka, M. Tasche, Invertible integer DCT algorithms, J.Sci. Signal
Processing: Image Communication, 15 (2003) 70-88.
[17] G. Plonka, A global method for inversible integer DCT and integer
wavelet algorithms, J.Sci.