Morphological Residue Encoding and Piecewise Approximation Techniques for Lossless Binary Image Compression

In this work, we propose an advanced framework for lossless compression of binary images. The proposed framework is organized in two steps: contour approximation and residue encoding. First, object contours are approximated by piecewise cubic polynomial curves and Improved Adaptive Arithmetic Coding are applied to encode the coefficients. Then, we propose an efficient algorithm based on morphological operation to detect the residues, i.e. the points of the original image that are not in the reconstructed image. After encoding the residue by the proposed column by column method or the Chebyshev distance method, a very high compression ratio can be achieved.