Abstract Bayesian inference provides a powerful framework to optimally integrate statistically learned prior knowledge into numerous computer vision algorithms. While the Bayesian approach has been successfully applied in the Markov random field literature, the resulting combinatorial optimization problems have been commonly treated with rather inefficient and inexact general purpose optimization methods such as Simulated Annealing. An efficient method to compute the global optima of certain classes of cost functions defined on binary-valued variables is given by graph min-cuts. In this paper, we propose to reconsider the problem of statistical learning for Bayesian inference in the context of efficient optimization schemes. Specifically, we address the question: Which prior information may be learned while retaining the ability to apply Graph Cut optimization? We provide a framework to learn and impose prior knowledge on the distribution of pairs and triplets of labels. As an illustration, we demonstrate that one can optimally restore binary textures from very noisy images with runtimes on the order of a second while imposing hundreds of statistically learned constraints per pixel.


    Access

    Download


    Export, share and cite



    Title :

    Statistical Priors for Efficient Combinatorial Optimization Via Graph Cuts


    Contributors:


    Publication date :

    2006-01-01


    Size :

    12 pages





    Type of media :

    Article/Chapter (Book)


    Type of material :

    Electronic Resource


    Language :

    English




    Statistical Priors for Efficient Combinatorial Optimization Via Graph Cuts

    Cremers, D. / Grady, L. | British Library Conference Proceedings | 2006


    Corrected Laplacians: closer cuts and segmentation with shape priors

    Tolliver, D. / Miller, G.L. / Collins, R.T. | IEEE | 2005


    Visual Saliency with Statistical Priors

    Li, J. / Tian, Y. / Huang, T. | British Library Online Contents | 2014



    Efficient Graph Cuts for Multiclass Interactive Image Segmentation

    Lu, Fangfang / Fu, Zhouyu / Robles-Kelly, Antonio | Springer Verlag | 2007