Previous [1] [2] [3] [4] [5] [6] [7] [8] [9] [10]

Journal of Inforamtion Science and Engineering, Vol.17 No.4, pp.697-708 (July 2001)

A Nearly Optimum Generalized Multilevel Block
Truncation Coding Algorithm With a Fast Nonexhaustive
Search Based on a Mean Square Error Criterion

Chun-Hung Kuo and Chang-Fuu Chen*

Advanced Technology Center
Computer and Communications Research Lab.
Industrial Technology Research Institute
Hsinchu, Taiwan 310, R.O.C.
*Department of Computer and Communication Engineering
National Kaohsiung First University of
Science and Technology
Kaohsiung, Taiwan 811, R.O.C.

Block Truncation Coding (BTC), using a one-bit moment preservign quantizer to quantize a block, is a simple compression method; however, compressin with BTC cannot be used to obtain a sufficiently high quality image, especially an image with many edges. In this paper, a new generalized multilevel BTC algorithm, which arbitrarily specifies the nubmer of quantization levels, is developed. In the proposed algorithm, a nearly optimum quantizer is proposed based on a multiplication-free and non-exhaustive search such that the computation complexity is very low. The simulation results indicate that both the computation complexity and the reconstructed image quality obtained using the proposed multilevel BTC algorithm are better than those obtainable with other existing multilevel BTC algorithms.

Keywords: image compression, block truncation coding (BTC), quantization, nonexhaustive search, mean square error criterion

Full Text () Retrieve PDF document (200107_10.pdf)

Received December 2, 1998; revised May 17 & July 22, 1999; accepted August 23, 1999.
Communicated by Soo-Chang Pei.