Name
Abstract | ||
|---|---|---|
Due to the intrinsic long-tailed distribution of objects in the real world, we are unlikely to be able to train an object recognizer/detector with many visual examples for each category. We have to share visual knowledge between object categories to enable learning with few or no training examples. In this paper, we show that local object similarity information--statements that pairs of categories are similar or dissimilar--is a very useful cue to tie different categories to each other for effective knowledge transfer. The key insight: Given a set of object categories which are similar and a set of categories which are dissimilar, a good object model should respond more strongly to examples from similar categories than to examples from dissimilar categories. To exploit this category-dependent similarity regularization, we develop a regularized kernel machine algorithm to train kernel classifiers for categories with few or no training examples. We also adapt the state-of-the-art object detector to encode object similarity constraints. Our experiments on hundreds of categories from the Labelme dataset show that our regularized kernel classifiers can make significant improvement on object categorization. We also evaluate the improved object detector on the PASCAL VOC 2007 benchmark dataset. |
Year | DOI | Venue |
|---|---|---|
2013 | 10.1109/TPAMI.2013.58 | IEEE Trans. Pattern Anal. Mach. Intell. |
Keywords | Field | DocType |
visual knowledge,similar object categories,comparative object similarity,dissimilar object categories,object category,deformable part model,object similarity constraints,kernel classifier training,kernel machines,learning (artificial intelligence),intrinsic long-tailed object distribution,knowledge management,object detector,object recognizer,good object model,knowledge transfer,state-of-the-art object detector,improved object detector,local object similarity information,regularized kernel machine algorithm,svm,image classification,object categorization,pascal voc,sharing,object detection,object similarity constraint,improved object categorization,category-dependent similarity regularization,pascal voc 2007 benchmark dataset,training example,support vector machines,labelme dataset,detectors,kernel,visualization,learning artificial intelligence | Computer science,Artificial intelligence,Contextual image classification,Kernel (linear algebra),Computer vision,LabelMe,Categorization,Object detection,Pattern recognition,Support vector machine,Object model,Kernel method,Machine learning | Journal |
Volume | Issue | ISSN |
35 | 10 | 1939-3539 |
Citations | PageRank | References |
11 | 0.51 | 20 |
Authors | ||
3 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Gang Wang | 1 | 2869 | 135.49 |
| D. A. Forsyth | 2 | 9227 | 1138.80 |
| Derek Hoiem | 3 | 4998 | 302.66 |