Name
Abstract | ||
|---|---|---|
This paper presents a novel approach for visual scene modeling and classification, investigating the combined use of text modeling methods and local invariant features. Our work attempts to elucidate (1) whether a text-like bag-of-visterms representation (histogram of quantized local visual features) is suitable for scene (rather than object) classification, (2) whether some analogies between discrete scene representations and text documents exist, and (3) whether unsupervised, latent space models can be used both as feature extractors for the classification task and to discover patterns of visual co-occurrence. Using several data sets, we validate our approach, presenting and discussing experiments on each of these issues. We first show, with extensive experiments on binary and multi-class scene classification tasks using a 9,500-image data set, that the bag-of-visterms representation consistently outperforms classical scene classification approaches. In other data sets we show that our approach competes with or outperforms other recent, more complex, methods. We also show that Probabilistic Latent Semantic Analysis (PLSA) generates a compact scene representation, discriminative for accurate classification, and more robust than the bag-of-visterms representation when less labeled training data is available. Finally, through aspect-based image ranking experiments, we show the ability of PLSA to automatically extract visually meaningful scene patterns, making such representation useful for browsing image collections. |
Year | DOI | Venue |
|---|---|---|
2007 | 10.1109/TPAMI.2007.1155 | IEEE Trans. Pattern Anal. Mach. Intell. |
Keywords | Field | DocType |
classical scene classification approach,discrete scene representation,scene classifica- tion,multiclass scene classification task,compact scene representation,quantized local descriptors,accurate classification,meaningful scene pattern,index terms— image representation,thousand words,latent aspect modeling.,object recognition,visual scene modeling,classification task,bov representation,data mining,layout,detectors,image segmentation,text analysis,feature extraction,probability,image classification,construction industry,indexing terms,probabilistic latent semantic analysis | Computer vision,Histogram,Pattern recognition,Ranking,Computer science,Feature extraction,Image segmentation,Probabilistic latent semantic analysis,Artificial intelligence,Contextual image classification,Discriminative model,Cognitive neuroscience of visual object recognition | Journal |
Volume | Issue | ISSN |
29 | 9 | 0162-8828 |
Citations | PageRank | References |
115 | 3.83 | 32 |
Authors | ||
5 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Pedro Quelhas | 1 | 261 | 21.51 |
| Florent Monay | 2 | 593 | 31.43 |
| Jean-Marc Odobez | 3 | 518 | 31.95 |
| Daniel Gatica-Perez | 4 | 4182 | 276.74 |
| Tinne Tuytelaars | 5 | 10161 | 609.66 |