Block Truncation Coding with Huffman Coding

A simple and easy to implement block truncation coding (BTC) method employing a set of predefined bit planes for encoding the bit planes by the conventional BTC is proposed for image compression. The set of predefined bit planes is defined independently of the images to be encoded. The bit plane is matched for maximum correlation in the set of predefined bit planes. The bit rate is considerably reduced by encoding the index of the predefined bit planes using Huffman coding. Experimental results show that the proposed method provides acceptable visual quality reconstructed images with reduced bit rate and high peak signal to noise ratio (PSNR) when compared to the adaptive BTC (ABTC) method. Because of the edge preserving capability of the proposed method, it can be incorporated in hybrid image coding methods in real time applications such as medical image compression to achieve improved performance.