Fast connected-component labeling for binary hexagonal images

Although hexagonal images have attracted many attentions, there is almost no report on connected-component labeling algorithm for binary hexagonal images. This paper studies this problem for the first time and presents a fast connected-component labeling algorithm for binary hexagonal images. We analyze the connectivity of two different type foreground pixels with their processed pixels when an image is processed in the raster scan order, and give corresponding processing masks. For labeling binary hexagonal images, although we can process pixels one by one in the first scan as in most of labeling algorithms, we propose an efficient algorithm by processing pixels two by two. We show that by our proposed algorithm, for labeling a binary hexagonal image, the average number of times for checking the neighbor pixels for processing a foreground pixel will decrease, thus it leads to a more efficiently processing. Experimental results demonstrated that our proposed method is more efficient than the algorithm extended straightly from the fastest labeling algorithm for rectangle binary images.