Lossless compression of images employing two dimensional differentiators

A two dimensional differentiation approach is presented for the lossless compression of several classes of images. This technique is based on computing weighted differences between neighboring pixels. The performance of the new technique given here and the existing Joint Photographic Experts Group (JPEG) predictors approach are evaluated and compared for the lossless compression of different classes of images in several applications. The proposed approach yields better results compared with that obtained employing the lossless JPEG scheme. Arithmetic coding is used to code the residue sequence of both approaches.