A New Binary Image Representation: Logicodes

Using bincodes to represent binary images is shown to be very simple and storage-saving. Given a set of bincodes, this paper presents two improved codes, namely, the logicodes and the restricted logicodes to represent binary images. We first transform the given bincodes into a set of logical expressions. Then a minimization technique is employed to reduce the storage space required for these logical expressions, thus obtaining the logicodes, on which set operations can be applied directly. Further, we put some restrictions into these logicodes to make each resulting logicode, called the restricted logicode, representing a connected black block. Given 20 different-type real images, experimental results show that our logicodes (restricted logicodes) present a saving of 29% to 44% (12% to 34%) with respect to bincodes. When compared to Sarkar's method, except spending a little more space, our proposed codes do have three advantages: (1) it is easier to extract the related geometrical coordinates; (2) the bincodes can be used as direct input; i.e., they do compress the bincodes further; and (3) each restricted logicode represents a connected block block.