English

A two-layer Conditional Random Field for the classification of partially occluded objects

Computer Vision and Pattern Recognition 2013-09-16 v2

Abstract

Conditional Random Fields (CRF) are among the most popular techniques for image labelling because of their flexibility in modelling dependencies between the labels and the image features. This paper proposes a novel CRF-framework for image labeling problems which is capable to classify partially occluded objects. Our approach is evaluated on aerial near-vertical images as well as on urban street-view images and compared with another methods.

Keywords

Cite

@article{arxiv.1307.3043,
  title  = {A two-layer Conditional Random Field for the classification of partially occluded objects},
  author = {Sergey Kosov and Pushmeet Kohli and Franz Rottensteiner and Christian Heipke},
  journal= {arXiv preprint arXiv:1307.3043},
  year   = {2013}
}

Comments

Conference Submission

R2 v1 2026-06-22T00:49:33.906Z