%0 Conference Paper %1 Kumar2003 %A Kumar, Sanjiv %A Hebert, M. %B Computer Vision, 2003. Proceedings. Ninth IEEE International Conference on %D 2003 %K Bayes Corel Ising Markov based classification, classifiers, computer conditional constraints, contextual database, detection, discriminative field, fields, framework, generalized generative hidden hierarchical image images, interaction, interactions, kernel labeling, likelihood linear local man-made marginal maximum model model, models, natural neighborhood parameters, posterior posterior, probabilistic processes, random regions, rule, segmentation, sequences solution, structure text texture vision, %P 1150--1157 vol.2 %T Discriminative random fields: a discriminative framework for contextual interaction in classification %X In this work we present discriminative random fields (DRFs), a discriminative framework for the classification of image regions by incorporating neighborhood interactions in the labels as well as the observed data. The discriminative random fields offer several advantages over the conventional Markov random field (MRF) framework. First, the DRFs allow to relax the strong assumption of conditional independence of the observed data generally used in the MRF framework for tractability. This assumption is too restrictive for a large number of applications in vision. Second, the DRFs derive their classification power by exploiting the probabilistic discriminative models instead of the generative models used in the MRF framework. Finally, all the parameters in the DRF model are estimated simultaneously from the training data unlike the MRF framework where likelihood parameters are usually learned separately from the field parameters. We illustrate the advantages of the DRFs over the MRF framework in an application of man-made structure detection in natural images taken from the Corel database.

@inproceedings{Kumar2003, __markedentry = {[mozaher]}, abstract = {In this work we present discriminative random fields (DRFs), a discriminative framework for the classification of image regions by incorporating neighborhood interactions in the labels as well as the observed data. The discriminative random fields offer several advantages over the conventional Markov random field (MRF) framework. First, the DRFs allow to relax the strong assumption of conditional independence of the observed data generally used in the MRF framework for tractability. This assumption is too restrictive for a large number of applications in vision. Second, the DRFs derive their classification power by exploiting the probabilistic discriminative models instead of the generative models used in the MRF framework. Finally, all the parameters in the DRF model are estimated simultaneously from the training data unlike the MRF framework where likelihood parameters are usually learned separately from the field parameters. We illustrate the advantages of the DRFs over the MRF framework in an application of man-made structure detection in natural images taken from the Corel database.}, added-at = {2009-09-12T19:19:34.000+0200}, author = {Kumar, Sanjiv and Hebert, M.}, biburl = {https://www.bibsonomy.org/bibtex/22473fadfcff8d25cab6bfe1cab2c9c1e/mozaher}, booktitle = {Computer Vision, 2003. Proceedings. Ninth IEEE International Conference on}, file = {01238478.pdf:Kumar2003.pdf:PDF}, interhash = {f460ad9fb48f915fc762f1139a02f6d9}, intrahash = {2473fadfcff8d25cab6bfe1cab2c9c1e}, keywords = {Bayes Corel Ising Markov based classification, classifiers, computer conditional constraints, contextual database, detection, discriminative field, fields, framework, generalized generative hidden hierarchical image images, interaction, interactions, kernel labeling, likelihood linear local man-made marginal maximum model model, models, natural neighborhood parameters, posterior posterior, probabilistic processes, random regions, rule, segmentation, sequences solution, structure text texture vision,}, owner = {mozaher}, pages = {1150--1157 vol.2}, timestamp = {2009-09-12T19:19:40.000+0200}, title = {Discriminative random fields: a discriminative framework for contextual interaction in classification}, year = 2003 }