Tree Based Search Algorithm for Binary Image Compression

The demand for large-scale image data grows increasingly fast, resulting in a need for efficient image compression. This work aims at improving the lossless compression of binary images. To this end, we propose the use of a tree-based optimization algorithm, which searches for the best partitions of the input image into non-uniform blocks, and for the best combination of scan directions, with the goal of achieving more efficient compression of the resulting sequences of intervals between successive symbols of the same kind. The tree-based search algorithm searches for the best grid structure for adaptively partitioning the image into blocks of varying sizes. Extensive simulations of this search algorithm on various datasets demonstrated that we can achieve significantly higher compression on average than various standard binary image compression methods such as the JPEG 2000 and JBIG2 methods.