Compression of Binary Images by Hierachical Decomposition Based on the Vector Quantization

In this paper we address the new lossy coding method for digital transmission or storage of binary images using hierachical decomposition based on the vector quantization. Binary(white-and-black) images , especially engineering drawings and weather maps, are containing a lot of blank spaces in the entire image. From this aspect, the proposed method are employed two popular compression thechniques: JBIG and VQ. And we introduce node merging method which can be reduced a high computational complexity and ensured a good image quality for unbalanced high-order model codebook design. In this paper we present a binary image compression algorithm suitable for engineering drawing database. It is shown that high compression can be achieved by extracting structural information through hierarchical context modeling. The application considered in this paper is engineering drawing of gas pipeline map in the New York City by Brooklyn Union Gas company. The proposed approach is also attractive for fast browsing due to hierarchical structure of the compression algorithm.