Abstract | ||
---|---|---|
We propose hinge-loss Markov random fields (HLMRFs), a powerful class of continuous-valued graphical models, for high-level computer vision tasks. HL-MRFs are characterized by log-concave density functions, and are able to perform efficient, exact inference. Their templated hinge-loss potential functions naturally encode soft-valued logical rules. Using the declarative modeling language probabilistic soft logic, one can easily define HL-MRFs via familiar constructs from first-order logic. We apply HLMRFs to the task of activity detection, using principles of collective classification. Our model is simple, intuitive and interpretable. We evaluate our model on two datasets and show that it achieves significant lift over the low-level detectors. |
Year | DOI | Venue |
---|---|---|
2013 | 10.1109/CVPRW.2013.157 | CVPR Workshops |
Keywords | Field | DocType |
templated hinge-loss,continuous-valued graphical model,exact inference,soft logic,collective classification,hinge-loss markov random fields,collective activity detection,hinge-loss markov random field,encode soft-valued logical rule,activity detection,declarative modeling language probabilistic,first-order logic,random processes,probabilistic logic,first order logic,accuracy,cognition,detectors,image classification,computer vision,markov processes,computational modeling | Hinge loss,Markov process,Computer science,Probabilistic CTL,Theoretical computer science,Artificial intelligence,Probabilistic logic,Computer vision,Pattern recognition,Markov model,Markov chain,Variable-order Markov model,Graphical model | Conference |
Volume | Issue | ISSN |
2013 | 1 | 2160-7508 |
Citations | PageRank | References |
6 | 0.42 | 24 |
Authors | ||
6 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ben London | 1 | 77 | 7.01 |
Sameh Khamis | 2 | 170 | 7.54 |
Stephen H. Bach | 3 | 66 | 5.70 |
Bert Huang | 4 | 563 | 39.09 |
Lise Getoor | 5 | 4365 | 320.21 |
Larry S. Davis | 6 | 14201 | 2690.83 |