Name
Abstract | ||
|---|---|---|
We present a novel approach for contextual segmentation of complex visual scenes, based on the use of bags of local invariant features (visterms) and probabilistic aspect mo d- els. Our approach uses context in two ways: (1) by using the fact that specific learned aspects correlate with the semant ic classes, which resolves some cases of visual polysemy, and (2) by formalizing the notion that scene context is image- specific -what an individual visterm represents depends on what the rest of the visterms in the same bag represent too-. We demonstrate the validity of our approach on a man-made vs. natural visterm classification problem. Experiments on an image collection of complex scenes show that the ap- proach improves region discrimination, producing satisfa c- tory results, and outperforming a non-contextual method. Furthermore, through the later use of a Markov Random Field model, we also show that co-occurrence and spatial contextual information can be conveniently integrated for improved visterm classification. |
Year | DOI | Venue |
|---|---|---|
2006 | 10.1109/CVPRW.2006.102 | CVPRW '06 Proceedings of the 2006 Conference on Computer Vision and Pattern Recognition Workshop |
Keywords | Field | DocType |
patchbased scene segmentation,spatial contextual information,natural visterm classification problem,novel approach,integrating co-occurrence,complex scene,individual visterm,complex visual scene,scene context,later use,contextual segmentation,improved visterm classification,spatial contexts,layout,computer vision,image segmentation,image retrieval,vision,pixel,spatial context,context modeling,indexing,image analysis,image resolution | Computer vision,Pattern recognition,Computer science,Markov random field,Segmentation,Search engine indexing,Image retrieval,Context model,Image segmentation,Artificial intelligence,Probabilistic logic,Polysemy | Conference |
Volume | Issue | ISSN |
2006 | 1 | 2160-7508 |
ISBN | Citations | PageRank |
0-7695-2646-2 | 8 | 0.83 |
References | Authors | |
15 | 4 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Florent Monay | 1 | 593 | 31.43 |
| Pedro Quelhas | 2 | 261 | 21.51 |
| Jean-marc Odobez | 3 | 1641 | 110.52 |
| Daniel Gatica-Perez | 4 | 4182 | 276.74 |