ELS Dunn 47-ch35-569-582-9780123849885 — 2011/1/11 — 15:54 — page 569 — #1 CHAPTER 35Connected Component Labelingin CUDA

Connected component labeling (CCL) is a task of detecting connected regions in input data, and it finds its applications in pattern recognition, computer vision, and image processing. We present a new algorithm for connected component labeling in 2-D images implemented in CUDA. We first provide a brief overview of the CCL problem together with existing CPU-oriented algorithms. The rest of the chapter is dedicated to a description of our CUDA CCL method. The goal of a CCL algorithm is to find a unique label for every set of connected elements in input data. In other words, CCL is used to divide input elements, for example, pixels of a raster image, into groups where all elements from a single group represent a connected object. CCL can be thought of as topological clustering based on connectivity. CCL is often confused with segmentation, which i