Inference methods for CRFs with co-occurrence statistics

DOI10.1007/s11263-012-0583-yzbMath1270.68350MaRDI QIDQ362141

L'ubor Ladický, Chris Russell, Philip H. S. Torr, Pushmeet Kohli

Publication date: 20 August 2013

Published in: International Journal of Computer Vision (Search for Journal in Brave)

Full work available at URL: https://doi.org/10.1007/s11263-012-0583-y

zbMATH Keywords

optimization; conditional random fields; object class segmentation

Mathematics Subject Classification ID

62M40: Random fields; image analysis

94A08: Image processing (compression, reconstruction, etc.) in information and communication theory

68T45: Machine vision and scene understanding

Related Items

Active-set Methods for Submodular Minimization Problems, Efficient Relaxations for Dense CRFs with Sparse Higher-Order Potentials, Fast approximate energy minimization with label costs, Generalized roof duality, Filter-based mean-field inference for random fields with higher-order terms and product label-spaces, Conditional random fields for pattern recognition applied to structured data, Integrating geometrical context for semantic labeling of indoor scenes using RGBD images, Midrange geometric interactions for semantic segmentation. Constraints for continuous multi-label optimization, Minimizing energies with hierarchical costs, Optimizing the Relevance-Redundancy Tradeoff for Efficient Semantic Segmentation

Cites Work

• Unnamed Item
• Fast approximate energy minimization with label costs
• An exact primal-dual penalty method approach to warmstarting interior-point methods for linear programming
• Approximation algorithms for classification problems with pairwise relationships
• MAP Estimation Via Agreement on Trees: Message-Passing and Linear Programming
• Geometric motion segmentation and model selection
• On the optimality of solutions of the max-product belief-propagation algorithm in arbitrary graphs