Generating fuzzy semantic metadata describing spatial relations from images using the R-Histogram

Yuhang Wang, Fillia Makedon, James Ford, Li Shen, Dina Goldin

Research output: Chapter in Book/Report/Conference proceedingConference contribution

19 Citations (Scopus)

Abstract

Automatic generation of semantic metadata describing spatial relations is highly desirable for image digital libraries. Relative spatial relations between objects in an image convey important information about the image. Because the perception of spatial relations is subjective, we propose a novel framework for automatic metadata generation based on fuzzy k-NN classification that generates fuzzy semantic metadata describing spatial relations between objects in an image. For each pair of objects of interest, the corresponding R-Histogram is computed and used as input for a set of fuzzy k-NN classifiers. The R-Histogram is a quantitative representation of spatial relations between two objects. The outputs of the classifiers are soft class labels for each of the following eight spatial relations: 1) LEFT OF, 2) RIGHT OF, 3) ABOVE, 4) BELOW, 5) NEAR, 6) FAR, 7) INSIDE, 8) OUTSIDE. Because the classifier-training stage involves annotating the training images manually, it is desirable to use as few training images as possible. To address this issue, we applied existing prototype selection techniques and also devised two new extensions. We evaluated the performance of different fuzzy k-NN algorithms and prototype selection algorithms empirically on both synthetic and real images. Preliminary experimental results show that our system is able to obtain good annotation accuracy (92%-98% on synthetic images and 82%-93% on real images) using only a small training set (4-5 images).

Original languageEnglish (US)
Title of host publicationProceedings of the ACM IEEE International Conference on Digital Libraries, JCDL 2004
EditorsH. Chen, M. Christel, E.P. Lim
Pages202-211
Number of pages10
StatePublished - 2004
Externally publishedYes
EventProceedings of the Fourth ACM/IEEE Joint Conference on Digital Libraries; Global reach and Diverse Impact, JCDL 2004 - Tucson, AZ, United States
Duration: Jun 7 2004Jun 11 2004

Other

OtherProceedings of the Fourth ACM/IEEE Joint Conference on Digital Libraries; Global reach and Diverse Impact, JCDL 2004
CountryUnited States
CityTucson, AZ
Period6/7/046/11/04

Fingerprint

Metadata
Classifiers
Semantics
Digital libraries
Labels

Keywords

  • Image Digital Library
  • k-Nearest Neighbor Rule
  • Metadata
  • Prototype Selection
  • R-Histogram
  • Spatial Relations

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Wang, Y., Makedon, F., Ford, J., Shen, L., & Goldin, D. (2004). Generating fuzzy semantic metadata describing spatial relations from images using the R-Histogram. In H. Chen, M. Christel, & E. P. Lim (Eds.), Proceedings of the ACM IEEE International Conference on Digital Libraries, JCDL 2004 (pp. 202-211)

Generating fuzzy semantic metadata describing spatial relations from images using the R-Histogram. / Wang, Yuhang; Makedon, Fillia; Ford, James; Shen, Li; Goldin, Dina.

Proceedings of the ACM IEEE International Conference on Digital Libraries, JCDL 2004. ed. / H. Chen; M. Christel; E.P. Lim. 2004. p. 202-211.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Wang, Y, Makedon, F, Ford, J, Shen, L & Goldin, D 2004, Generating fuzzy semantic metadata describing spatial relations from images using the R-Histogram. in H Chen, M Christel & EP Lim (eds), Proceedings of the ACM IEEE International Conference on Digital Libraries, JCDL 2004. pp. 202-211, Proceedings of the Fourth ACM/IEEE Joint Conference on Digital Libraries; Global reach and Diverse Impact, JCDL 2004, Tucson, AZ, United States, 6/7/04.
Wang Y, Makedon F, Ford J, Shen L, Goldin D. Generating fuzzy semantic metadata describing spatial relations from images using the R-Histogram. In Chen H, Christel M, Lim EP, editors, Proceedings of the ACM IEEE International Conference on Digital Libraries, JCDL 2004. 2004. p. 202-211
Wang, Yuhang ; Makedon, Fillia ; Ford, James ; Shen, Li ; Goldin, Dina. / Generating fuzzy semantic metadata describing spatial relations from images using the R-Histogram. Proceedings of the ACM IEEE International Conference on Digital Libraries, JCDL 2004. editor / H. Chen ; M. Christel ; E.P. Lim. 2004. pp. 202-211
@inproceedings{c9e8da4e0b8d4725be26d1635ea3ae39,
title = "Generating fuzzy semantic metadata describing spatial relations from images using the R-Histogram",
abstract = "Automatic generation of semantic metadata describing spatial relations is highly desirable for image digital libraries. Relative spatial relations between objects in an image convey important information about the image. Because the perception of spatial relations is subjective, we propose a novel framework for automatic metadata generation based on fuzzy k-NN classification that generates fuzzy semantic metadata describing spatial relations between objects in an image. For each pair of objects of interest, the corresponding R-Histogram is computed and used as input for a set of fuzzy k-NN classifiers. The R-Histogram is a quantitative representation of spatial relations between two objects. The outputs of the classifiers are soft class labels for each of the following eight spatial relations: 1) LEFT OF, 2) RIGHT OF, 3) ABOVE, 4) BELOW, 5) NEAR, 6) FAR, 7) INSIDE, 8) OUTSIDE. Because the classifier-training stage involves annotating the training images manually, it is desirable to use as few training images as possible. To address this issue, we applied existing prototype selection techniques and also devised two new extensions. We evaluated the performance of different fuzzy k-NN algorithms and prototype selection algorithms empirically on both synthetic and real images. Preliminary experimental results show that our system is able to obtain good annotation accuracy (92{\%}-98{\%} on synthetic images and 82{\%}-93{\%} on real images) using only a small training set (4-5 images).",
keywords = "Image Digital Library, k-Nearest Neighbor Rule, Metadata, Prototype Selection, R-Histogram, Spatial Relations",
author = "Yuhang Wang and Fillia Makedon and James Ford and Li Shen and Dina Goldin",
year = "2004",
language = "English (US)",
isbn = "1581138326",
pages = "202--211",
editor = "H. Chen and M. Christel and E.P. Lim",
booktitle = "Proceedings of the ACM IEEE International Conference on Digital Libraries, JCDL 2004",

}

TY - GEN

T1 - Generating fuzzy semantic metadata describing spatial relations from images using the R-Histogram

AU - Wang, Yuhang

AU - Makedon, Fillia

AU - Ford, James

AU - Shen, Li

AU - Goldin, Dina

PY - 2004

Y1 - 2004

N2 - Automatic generation of semantic metadata describing spatial relations is highly desirable for image digital libraries. Relative spatial relations between objects in an image convey important information about the image. Because the perception of spatial relations is subjective, we propose a novel framework for automatic metadata generation based on fuzzy k-NN classification that generates fuzzy semantic metadata describing spatial relations between objects in an image. For each pair of objects of interest, the corresponding R-Histogram is computed and used as input for a set of fuzzy k-NN classifiers. The R-Histogram is a quantitative representation of spatial relations between two objects. The outputs of the classifiers are soft class labels for each of the following eight spatial relations: 1) LEFT OF, 2) RIGHT OF, 3) ABOVE, 4) BELOW, 5) NEAR, 6) FAR, 7) INSIDE, 8) OUTSIDE. Because the classifier-training stage involves annotating the training images manually, it is desirable to use as few training images as possible. To address this issue, we applied existing prototype selection techniques and also devised two new extensions. We evaluated the performance of different fuzzy k-NN algorithms and prototype selection algorithms empirically on both synthetic and real images. Preliminary experimental results show that our system is able to obtain good annotation accuracy (92%-98% on synthetic images and 82%-93% on real images) using only a small training set (4-5 images).

AB - Automatic generation of semantic metadata describing spatial relations is highly desirable for image digital libraries. Relative spatial relations between objects in an image convey important information about the image. Because the perception of spatial relations is subjective, we propose a novel framework for automatic metadata generation based on fuzzy k-NN classification that generates fuzzy semantic metadata describing spatial relations between objects in an image. For each pair of objects of interest, the corresponding R-Histogram is computed and used as input for a set of fuzzy k-NN classifiers. The R-Histogram is a quantitative representation of spatial relations between two objects. The outputs of the classifiers are soft class labels for each of the following eight spatial relations: 1) LEFT OF, 2) RIGHT OF, 3) ABOVE, 4) BELOW, 5) NEAR, 6) FAR, 7) INSIDE, 8) OUTSIDE. Because the classifier-training stage involves annotating the training images manually, it is desirable to use as few training images as possible. To address this issue, we applied existing prototype selection techniques and also devised two new extensions. We evaluated the performance of different fuzzy k-NN algorithms and prototype selection algorithms empirically on both synthetic and real images. Preliminary experimental results show that our system is able to obtain good annotation accuracy (92%-98% on synthetic images and 82%-93% on real images) using only a small training set (4-5 images).

KW - Image Digital Library

KW - k-Nearest Neighbor Rule

KW - Metadata

KW - Prototype Selection

KW - R-Histogram

KW - Spatial Relations

UR - http://www.scopus.com/inward/record.url?scp=4944235921&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=4944235921&partnerID=8YFLogxK

M3 - Conference contribution

SN - 1581138326

SN - 9781581138320

SP - 202

EP - 211

BT - Proceedings of the ACM IEEE International Conference on Digital Libraries, JCDL 2004

A2 - Chen, H.

A2 - Christel, M.

A2 - Lim, E.P.

ER -