2012 IEEE Conference on Computer Vision and Pattern Recognition (2012)
Providence, RI USA
June 16, 2012 to June 21, 2012
Xiaodong Wu , Dept. of Electr. & Comput. Eng., Univ. of Iowa, Iowa City, IA, USA
O. Veksler , Comput. Sci. Dept., Univ. of Western Ontario, London, ON, Canada
Qi Song , Dept. of Electr. & Comput. Eng., Univ. of Iowa, Iowa City, IA, USA
Junjie Bai , Dept. of Electr. & Comput. Eng., Univ. of Iowa, Iowa City, IA, USA
Many computer vision applications can be formulated as labeling problems. However, multilabeling problems are usually very challenging to solve, especially when some ordering constraints are enforced. We solve in this paper a five-parts labeling problem proposed in [6, 7]. In this model, one wants to find an optimal labeling for an image with five possible parts: “left”, “right”, “top”, “bottom” and “center”. The geometric ordering constraints can be read naturally from the names. No previous method can solve the problem with globally optimal solutions in a linear space complexity. We propose an efficient dynamic programming based algorithm which guarantees the global optimal labeling for the five-parts model. The time complexity is O(N1.5) and the space complexity is O(N), with N being the number of pixels in the image. In practice, it runs faster than previous methods. Moreover, it works for both 4-neighborhood and 8-neighborhood settings, and can be easily parallelized for GPU.
graphics processing units, computational complexity, computer vision, constraint handling, dynamic programming, geometry, GPU, computer vision applications, multilabeling problems, five-part labeling problem, geometric ordering constraints, linear space complexity, dynamic programming based algorithm, global optimal image labeling, five-part model, image left parts, image right parts, image top, image bottom, image center, globally optimal solutions, time complexity, Labeling, Heuristic algorithms, Complexity theory, Green products, Optimization, Memory management, Dynamic programming
Xiaodong Wu, O. Veksler, Qi Song and Junjie Bai, "Fast dynamic programming for labeling problems with ordering constraints," 2012 IEEE Conference on Computer Vision and Pattern Recognition(CVPR), Providence, RI USA, 2012, pp. 1728-1735.