Beyond Bags of Features: Spatial Pyramid Matching for Recognizing Natural Scene Categories

by: Svetlana Lazebnik, Cordelia Schmid, Jean Ponce

(2006), pp. 2169-2178.

View FullText article

ACM, DOI, IEEE Explore, http://www.vision.caltech.edu/Image_Datas…

Reviews [Write a review of this article]

There are no reviews of this article

Find related articles from these CiteULike users

iansimon, vinaypn, dlmace, jccaicedo, mamadoudiao, zeppe, GernotMargreitner, caicai

Find related articles with these CiteULike tags

classification, detection, feature, features, generals, image, lazebnik06, matching, no-tag, object, plsa, recognition, retrieval, segmentation, spatial, vision, visual-bow

Abstract

This paper presents a method for recognizing scene categories based on approximate global geometric correspondence. This technique works by partitioning the image into increasingly fine sub-regions and computing histograms of local features found inside each sub-region. The resulting "spatial pyramid" is a simple and computationally efficient extension of an orderless bag-of-features image representation, and it shows significantly improved performance on challenging scene categorization tasks. Specifically, our proposed method exceeds the state of the art on the Caltech-101 database and achieves high accuracy on a large database of fifteen natural scene categories. The spatial pyramid framework also offers insights into the success of several recently proposed image descriptions, including Torralba's "gist" and Lowe's SIFT descriptors.

@inproceedings{citeulike:1720832, abstract = {This paper presents a method for recognizing scene categories based on approximate global geometric correspondence. This technique works by partitioning the image into increasingly fine sub-regions and computing histograms of local features found inside each sub-region. The resulting "spatial pyramid" is a simple and computationally efficient extension of an orderless bag-of-features image representation, and it shows significantly improved performance on challenging scene categorization tasks. Specifically, our proposed method exceeds the state of the art on the Caltech-101 database and achieves high accuracy on a large database of fifteen natural scene categories. The spatial pyramid framework also offers insights into the success of several recently proposed image descriptions, including Torralba's "gist" and Lowe's SIFT descriptors.}, address = {Washington, DC, USA}, author = {Lazebnik, Svetlana and Schmid, Cordelia and Ponce, Jean }, booktitle = {CVPR '06: Proceedings of the 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition}, citeulike-article-id = {1720832}, doi = {10.1109/CVPR.2006.68}, pages = {2169--2178}, posted-at = {2008-05-06 18:05:49}, priority = {0}, publisher = {IEEE Computer Society}, title = {Beyond Bags of Features: Spatial Pyramid Matching for Recognizing Natural Scene Categories}, url = {http://portal.acm.org/citation.cfm?id=1153549}, year = {2006} }

RIS record

TY - CONF ID - citeulike:1720832 L3 - citeulike-article-id:1720832 TI - Beyond Bags of Features: Spatial Pyramid Matching for Recognizing Natural Scene Categories T2 - CVPR '06: Proceedings of the 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition SP - 2169 EP - 2178 AD - Washington, DC, USA PB - IEEE Computer Society N2 - This paper presents a method for recognizing scene categories based on approximate global geometric correspondence. This technique works by partitioning the image into increasingly fine sub-regions and computing histograms of local features found inside each sub-region. The resulting "spatial pyramid" is a simple and computationally efficient extension of an orderless bag-of-features image representation, and it shows significantly improved performance on challenging scene categorization tasks. Specifically, our proposed method exceeds the state of the art on the Caltech-101 database and achieves high accuracy on a large database of fifteen natural scene categories. The spatial pyramid framework also offers insights into the success of several recently proposed image descriptions, including Torralba's "gist" and Lowe's SIFT descriptors. AU - Lazebnik, Svetlana AU - Schmid, Cordelia AU - Ponce, Jean PY - 2006/// UR - http://portal.acm.org/citation.cfm?id=1153549 ER -

RIS BibTeX