A fast region expansion labeling of connected components in binary image

In this paper, we propose a new image labeling algorithm for object analysis of the binary images. With one-scan process, the foreground (object) pixel is assigned a provisional label, and label equivalences between provisional labels are merged later. Our approach merges the region of the labeling of the foreground image during the one-scan labeling process. Compared with the existing conventional labeling algorithms, experimental results verify that our algorithm labels the rectangle-shaped objects with lower complexity.