Lossless contour compression using morphology, chain coding, and distribution transform

Abstract

Chain coding is widely used in image compression to encode the boundaries of objects efficiently. Although chain codes are effective, they still need large amount of memory to store the codes. Therefore, an efficient encoding technique for chain codes is required. In this paper, we propose an algorithm to encode contours losslessly. First, the morphological operation is applied to shrink the contours if the process is invertible. Then, the modified Angle Freeman chain code is used to represent the contours, and the distribution transform is applied to rearrange the binary stream and the proposed improved adaptive arithmetic code is adopted for encoding. Simulations show that the proposed algorithm can much reduce the data size required for encoding contours.