A quantization based codebook formation method of vector quantization algorithm to improve the compression ratio while preserving the visual quality of the decompressed image

An extremely difficult problem in an image compression method is increasing the compression ratio while maintaining the visual quality of the image. One of the popular imae compression strategies that can be found in literature is vector quantization. The size of the Codebook and Index matrix created as a result of this procedure determines how effective it is. The Codebook has a size of nc*cwd, where cwd is the codeword dimension and nc is the number of clusters. These n number of codewords do not all have an equal impact on the visual quality of the decompressed image. Some of these are essential for rebuilding the decompressed image, while others are less significant. In this paper, a novel strategy is put forth to divide a codebook's nc number of codewords into two groups. The codewords used most frequently in reconstructing the image are in one section, while the rest are included in the second group of less crucial codewords. The first group’s codewords are left in tact to preserve the decompressed image’s visual quality, but the second group’s codewords are quantized into two bit values, namely 0, 1, 2, and 3, to increase compression ratio. The suggested technique is used on numerous colour images from the UCIDv.2 database as well as images from the standard image database. Peak signal to noise ratio, structural similarity ındex measure and Compression Ratio are used to analyse the experimental outcomes. Depending on the size of the original image, experimental results reveal that the suggested method reduces the codebook size by 32.01–54.80% while maintaining the quality of the decompressed image. This greatly improves the compression ratio of the algorithm.