An efficient image compression algorithm for almost dual-color image based on k-means clustering, bit-map generation and RLE

In this paper, we have proposed an image compression/decompression algorithm for almost dual-colour images using k-means clustering, bit-map generation and run length encoding. The proposed algorithm is very simple in implementation, fast in encoding time. Experimental results show that compression ratio of this algorithm is better than our previous BOBC algorithm, dynamic BOBC-RLE algorithm and JPEG compression techniques. Image quality (PSNR) better as compared to that of the above mentioned compression techniques. In actuality the algorithm not only compresses the image but also sharpens it and removes color artefacts that result at the border between two colors.