A new noiseless coding technique for binary images

In this paper, a new coding technique for binary images is proposed. This technique makes use of the fact that boundaries contain all the information about such images. The boundary information is obtained by consecutive application of a simple one-dimensional technique in each of the spatial dimensions. The compression efficiency is then further increased by the reduction of correlation between co-ordinates of the coded pixels. Some sample images of varying complexity and type are used to compare the performances of the proposed technique and of other previously introduced schemes. The comparison shows that the proposed technique is generally more efficient than other techniques, especially for images of high complexity.