Classifying image galleries into a taxonomy using metadata and wikipedia

Authors:
Gerwin Kramer;Gosse Bouma;Dennis Hendriksen;Mathijs Homminga
Affiliations:
Information Science, University of Groningen, The Netherlands;Information Science, University of Groningen, The Netherlands;Kalooga, Groningen, The Netherlands;Kalooga, Groningen, The Netherlands
Venue:
NLDB'12 Proceedings of the 17th international conference on Applications of Natural Language Processing and Information Systems
Year:
2012

Citing 8
Cited 1

Does ontology help in image retrieval?: a comparison between keyword, text ontology and multi-modality ontology approaches

MULTIMEDIA '06 Proceedings of the 14th annual ACM international conference on Multimedia
Wikify!: linking documents to encyclopedic knowledge

Proceedings of the sixteenth ACM conference on Conference on information and knowledge management
Learning to link with wikipedia

Proceedings of the 17th ACM conference on Information and knowledge management
Using Wikipedia knowledge to improve text classification

Knowledge and Information Systems
Mining meaning from Wikipedia

International Journal of Human-Computer Studies
Deriving a large scale taxonomy from Wikipedia

AAAI'07 Proceedings of the 22nd national conference on Artificial intelligence - Volume 2
SemanticRank: ranking keywords and sentences using semantic graphs

COLING '10 Proceedings of the 23rd International Conference on Computational Linguistics
Robust disambiguation of named entities in text

EMNLP '11 Proceedings of the Conference on Empirical Methods in Natural Language Processing

Editorial: Minimally-supervised learning of domain-specific causal relations using an open-domain corpus as knowledge base

Data & Knowledge Engineering

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper presents a method for the hierarchical classification of image galleries into a taxonomy. The proposed method links textual gallery metadata to Wikipedia pages and categories. Entity extraction from metadata, entity ranking, and selection of categories is based on Wikipedia and does not require labeled training data. The resulting system performs well above a random baseline, and achieves a (micro-averaged) F-score of 0.59 on the 9 top categories of the taxonomy and 0.40 when using all 57 categories.