The Robotics Institute

Carnegie Mellon Robotics Institute

Copyright notice: This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder.

Abstract

Estimation of parameters of random field models from la- beled training data is crucial for their good performance in many im- age analysis applications. In this paper, we present an approach for ap- proximate maximum likelihood parameter learning in discriminative field models, which is based on approximating true expectations with simple piecewise constant functions constructed using inference techniques. Gra- dient ascent with these updates exhibits compelling limit cycle behavior which is tied closely to the number of errors made during inference. The performance of various approximations was evaluated with different in- ference techniques showing that the learned parameters lead to good classification performance so long as the method used for approximating the gradient is consistent with the inference mechanism. The proposed approach is general enough to be used for the training of, e.g., smoothing parameters of conventional Markov Random Fields (MRFs).

Notes

Associated Center(s) / Consortia: Vision and Autonomous Systems Center

Text Reference

Sanjiv Kumar, Jonas August, and Martial Hebert, "Exploiting Inference for Approximate Parameter Learning in Discriminative Fields: An Empirical Study," Energy Minimization Methods in Computer Vision and Pattern Recognition (EMMCVPR), 2005.

BibTeX Reference

@inproceedings{Kumar_2005_5281,
   author = "Sanjiv Kumar and Jonas August and Martial Hebert",
   title = "Exploiting Inference for Approximate Parameter Learning in Discriminative Fields: An Empirical Study",
   booktitle = "Energy Minimization Methods in Computer Vision and Pattern Recognition (EMMCVPR)",
   year = "2005",
}

The Robotics Institute

Menu