A novel color image coding technique using improved BTC with k-means quad clustering

Abstract

A new approach to color image compression on HSV model with good quality of reconstructed images and better compression ratio using Improved Block Truncation Coding algorithm with k-means Quad Clustering (IBTC-KQ) is proposed in this paper. The RGB plane of the color image is transformed into HSV plane in order to reduce the high degree of correlation between the RGB planes. Each HSV plane is then encoded using IBTC-KQ method. The block sizes are chosen based on the information content of the respective plane. The result of the proposed method is compared with that of other BTC based methods on RGB model and it shows a better performance both in the visual quality and compression ratio. Also the proposed method involves only less number of simple computations when compared with other BTC methods. © Springer International Publishing Switzerland 2014.