Lossless compression of images using a hybrid transform/differentiation technique

A technique for the lossless compression of images is proposed. The algorithm employed here is based on a hybrid system utilizing a one dimensional (1-D) discrete cosine transform (DCT) to decorrelate the adjacent pixels in each image line in the spatial domain and a one dimensional differentiator to reduce the redundancy in the adjacent transform points in each line in the perpendicular direction. The proposed lossless technique is developed and applied to different classes of images. Its performance is compared, for these applications, with the two dimensional (2-D) lossless DCT approach. Simulation results indicate that the proposed technique provides increased compression while requiring less computations than the lossless 2-D DCT scheme.