Block truncation coding with entropy coding

Abstract

Block Truncation Coding (BTC) is a simple and fast image compression algorithm wich achieves constant bit rate of 2.0 bits per pixel. The method is however suboptimal. In the present paper we propose a modification of BTC in which the compression ratio will be improved by coding the quantization data and the bit plane by arithmetic coding with an adaptive modelling scheme. The results compare favourable with other BTC variants. The bit rate for test image Lena is 1.53 bits per pixel with the mean square error of 16.51.

Fast Image Compression by Quadtree Prediction

“…It is simpler to use just pixel values than larger blocks, because there may be many neighboring leaf blocks (of different sizes) above and to the left of the current block. Other predictors, which we found to be somewhat less precise, were suggested in [9,17].…”

“…This is rather common in progressive transmission methods; in lossless coding, knowing the (precise) parent intensity enables computing one of the four children, given the other three [4]. A prediction scheme based on neighbour blocks was suggested in [17] for block truncation coding. In [14], prediction was based on both the parent and neighbors.…”

Bit rate reduction techniques for absolute moment block truncation coding

“…For a 512x512 input image with 4x4 blocks, The size of a subimage is 128x128. Several methods for coding the quantization data have been proposed in the literature, including the fixed number of bits, vector quantization (Udpikar and Raina, 1987), DCT (Wu and Coll, 1991), and lossless coding (Franti and Nevalainen, 1995). The subimages have details and important features that must be preserved.…”