Title :
A Scalable graph-cut algorithm for N-D grids
Author :
Delong, Andrew ; Boykov, Yuri
Author_Institution :
Univ. of Western Ontario, London, ON
Abstract :
Global optimisation via s-t graph cuts is widely used in computer vision and graphics. To obtain high-resolution output, graph cut methods must construct massive N-D grid-graphs containing billions of vertices. We show that when these graphs do not fit into physical memory, current max-flow/min-cut algorithms-the workhorse of graph cut methods-are totally impractical. Others have resorted to banded or hierarchical approximation methods that get trapped in local minima, which loses the main benefit of global optimisation. We enhance the push-relabel algorithm for maximum flow [14] with two practical contributions. First, true global minima can now be computed on immense grid-like graphs too large for physical memory. These graphs are ubiquitous in computer vision, medical imaging and graphics. Second, for commodity multi-core platforms our algorithm attains near-linear speedup with respect to number of processors. To achieve these goals, we generalised the standard relabeling operations associated with push-relabel.
Keywords :
computer vision; graph theory; optimisation; N-D grids; approximation methods; computer vision; global optimisation; graph-cut algorithm; graphics; immense grid-like graphs; local minima; max-flow/min-cut algorithms; maximum flow; medical imaging; multicore platforms; push-relabel algorithm; s-t graph cuts; Approximation methods; Biomedical imaging; Computer graphics; Computer vision; Grid computing; Image reconstruction; Image segmentation; Optimization methods; Pervasive computing; Physics computing;
Conference_Titel :
Computer Vision and Pattern Recognition, 2008. CVPR 2008. IEEE Conference on
Conference_Location :
Anchorage, AK
Print_ISBN :
978-1-4244-2242-5
Electronic_ISBN :
1063-6919
DOI :
10.1109/CVPR.2008.4587464