Efficient Learning to Label Images

Date

2010

Authors

Jia, Ke
Cheng, Li
Liu, Nianjun
Wang, Lei

Journal Title

Journal ISSN

Volume Title

Publisher

Institute of Electrical and Electronics Engineers (IEEE Inc)

Abstract

Conditional random field methods (CRFs) have gained popularity for image labeling tasks in recent years. In this paper, we describe an alternative discriminative approach, by extending the large margin principle to incorporate spatial correlations among neighboring pixels. In particular, by explicitly enforcing the submodular condition, graph-cuts is conveniently integrated as the inference engine to attain the optimal label assignment efficiently. Our approach allows learning a model with thousands of parameters, and is shown to be capable of readily incorporating higher-order scene context. Empirical studies on a variety of image datasets suggest that our approach performs competitively compared to the state-of-the-art scene labeling methods.

Description

Keywords

Keywords: Conditional random field; Discriminative approach; Efficient learning; Empirical studies; Graph-cuts; Higher order; Image datasets; Image labeling; Label images; Labeling methods; Large margin principle; Spatial correlations; Submodular; Pattern recogniti

Citation

Source

Proceedings of The 23rd IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2010)

Type

Conference paper

Book Title

Entity type

Access Statement

License Rights

Restricted until

2037-12-31