A Bayesian Hierarchical Model for Learning Natural Scene Categories Export

by: Li Fei-Fei, Pietro Perona

In Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR '05), Vol. 2 (2005), pp. 524-531.

[Posts]

View FullText article

ACM, DOI, IEEE Explore, Rent at DeepDyve

bsilverthorn's tags for this article

bayesian image_modeling scene_categorization

Reviews [Write a review of this article]

Notes for this article

bsilverthorn has 1 private note and 0 public notes for this article. If you are bsilverthorn then you can log in to see the private note.

Find related articles from these CiteULike users

zhengyun, jrodriguez, fabgoos, klmoravec, rayzhang, bsilverthorn, jccaicedo, egozi, weiweiguo, pcarbo, iansimon, zeppe, caicai, julianpan

Find related articles with these CiteULike tags

annotation, bag-of-features, bayesian, c-bag-of-features, codebook, computer-vision, fei-fei05, graphical-models, image, image_categorization, image_classification, image-classification, image_modeling, image_processing, lda, local-image-descriptors, no-tag, object-categorization, object_models, object-recognition, part-based, recognition, scene_categorization, sceneseg, segmentation, seminar, s-pending, t-imaging, variational-methods

Abstract

We propose a novel approach to learn and recognize natural scene categories. Unlike previous work [9, 17], it does not require experts to annotate the training set. We represent the image of a scene by a collection of local regions, denoted as codewords obtained by unsupervised learning. Each region is represented as part of a "theme". In previous work, such themes were learnt from hand-annotations of experts, while our method learns the theme distributions as well as the codewords distribution over the themes without supervision. We report satisfactory categorization performances on a large set of 13 categories of complex scenes.

@inproceedings{citeulike:1401348, abstract = {We propose a novel approach to learn and recognize natural scene categories. Unlike previous work [9, 17], it does not require experts to annotate the training set. We represent the image of a scene by a collection of local regions, denoted as codewords obtained by unsupervised learning. Each region is represented as part of a "theme". In previous work, such themes were learnt from hand-annotations of experts, while our method learns the theme distributions as well as the codewords distribution over the themes without supervision. We report satisfactory categorization performances on a large set of 13 categories of complex scenes.}, address = {Washington, DC, USA}, author = {Fei-Fei, Li and Perona, Pietro}, booktitle = {Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR '05)}, citeulike-article-id = {1401348}, citeulike-linkout-0 = {http://portal.acm.org/citation.cfm?id=1068508.1069129}, citeulike-linkout-1 = {http://dx.doi.org/10.1109/CVPR.2005.16}, citeulike-linkout-2 = {http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1467486}, doi = {10.1109/CVPR.2005.16}, isbn = {0-7695-2372-2}, keywords = {bayesian, image\_modeling, scene\_categorization}, pages = {524--531}, posted-at = {2009-04-17 20:34:23}, priority = {0}, publisher = {IEEE Computer Society}, title = {{A Bayesian Hierarchical Model for Learning Natural Scene Categories}}, url = {http://dx.doi.org/10.1109/CVPR.2005.16}, volume = {2}, year = {2005} }

RIS record

TY - CONF ID - citeulike:1401348 L3 - citeulike-article-id:1401348 N2 - We propose a novel approach to learn and recognize natural scene categories. Unlike previous work [9, 17], it does not require experts to annotate the training set. We represent the image of a scene by a collection of local regions, denoted as codewords obtained by unsupervised learning. Each region is represented as part of a "theme". In previous work, such themes were learnt from hand-annotations of experts, while our method learns the theme distributions as well as the codewords distribution over the themes without supervision. We report satisfactory categorization performances on a large set of 13 categories of complex scenes. SN - 0-7695-2372-2 AD - Washington, DC, USA EP - 531 TI - {A Bayesian Hierarchical Model for Learning Natural Scene Categories} PB - IEEE Computer Society VL - 2 T2 - Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR '05) SP - 524 KW - bayesian KW - image_modeling KW - scene_categorization AU - Fei-Fei, Li AU - Perona, Pietro PY - 2005/// UR - http://dx.doi.org/10.1109/CVPR.2005.16 ER -

A Bayesian Hierarchical Model for Learning Natural Scene Categories Export

Citation Format