Boolean decomposition of binary image operators

Boolean expansion of binary functions has been used in the representation of binary image operators. This representation scheme induced the creation of image processing architectures based on decision diagrams. In some situations, the application of this graph-based operator over the input images is faster than the corresponding conventional approach, in that the operators are constructively described by means of basic operators and set operations. However, when the conventional architecture exploits the bit parallelism of the microprocessor logical instructions, its performance often surpasses that of the graph-based system. This article proposes a novel approach, in which the decision diagrams guide the actions of the underlying conventional architecture, taking advantage of its parallelism, so that faster image processing can be achieved.