Development of a near lossless image compression technique using mixed transforms

A lossless compression scheme for images is presented in this paper employing mixed transforms. First, the mixed transforms technique is applied to compress images in a lossy manner. The image is represented employing subsets of basis functions of two or more transforms. The coefficients are quantized and the image is reconstructed. The resulting reconstructed image samples are rounded to the nearest integer and the modified residual error is computed. This error is transmitted employing a lossless technique such as Huffman coding. It is shown that for a given number of retained coefficients, the mixed transform representation produces a smaller Root Mean Square Residual Error as compared to employing the DCT alone. Also, it is shown that the first order entropy of the error is smaller for the mixed transforms technique than for the DCT, thus resulting in smaller length Huffman codes.