Contextual Bag-of-Words for Visual Categorization

Cited 77 time in webofscience Cited 0 time in scopus
  • Hit : 527
  • Download : 793
Bag-of-words (BOW), which represents an image by the histogram of local patches on the basis of a visual vocabulary, has attracted intensive attention in visual categorization due to its good performance and flexibility. Conventional BOW neglects the contextual relations between local patches due to its Naive Bayesian assumption. However, it is well known that contextual relations play an important role for human beings to recognize visual categories from their local appearance. This paper proposes a novel contextual bag-of-words (CBOW) representation to model two kinds of typical contextual relations between local patches, i.e., a semantic conceptual relation and a spatial neighboring relation. To model the semantic conceptual relation, visual words are grouped on multiple semantic levels according to the similarity of class distribution induced by them, accordingly local patches are encoded and images are represented. To explore the spatial neighboring relation, an automatic term extraction technique is adopted to measure the confidence that neighboring visual words are relevant. Word groups with high relevance are used and their statistics are incorporated into the BOW representation. Classification is taken using the support vector machine with an efficient kernel to incorporate the relational information. The proposed approach is extensively evaluated on two kinds of visual categorization tasks, i.e., video event and scene categorization. Experimental results demonstrate the importance of contextual relations of local patches and the CBOW shows superior performance to conventional BOW.
Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
Issue Date
2011-04
Language
English
Article Type
Article
Keywords

OBJECT CATEGORIZATION; DISCOVERING OBJECTS; IMAGE FEATURES; CLASSIFICATION; SCENE; REPRESENTATION; DESCRIPTORS; RECOGNITION; MODELS; SHAPE

Citation

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, v.21, no.4, pp.381 - 392

ISSN
1051-8215
URI
http://hdl.handle.net/10203/21358
Appears in Collection
EE-Journal Papers(저널논문)
Files in This Item
000289212000001.pdf(825.95 kB)Download
This item is cited by other documents in WoS
⊙ Detail Information in WoSⓡ Click to see webofscience_button
⊙ Cited 77 items in WoS Click to see citing articles in records_button

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0