Image annotations by combining multiple evidence & wordNet Export

@inproceedings{citeulike:4010908, abstract = {The development of technology generates huge amounts of non-textual information, such as images. An efficient image annotation and retrieval system is highly desired. Clustering algorithms make it possible to represent visual features of images with finite symbols. Based on this, many statistical models, which analyze correspondence between visual features and words and discover hidden semantics, have been published. These models improve the annotation and retrieval of large image databases. However, current state of the art including our previous work produces too many irrelevant keywords for images during annotation. In this paper, we propose a novel approach that augments the classical model with generic knowledge-based, WordNet. Our novel approach strives to prune irrelevant keywords by the usage of WordNet. To identify irrelevant keywords, we investigate various semantic similarity measures between keywords and finally fuse outcomes of all these measures together to make a final decision using Dempster-Shafer evidence combination. We have implemented various models to link visual tokens with keywords based on knowledge-based, WordNet and evaluated performance using precision, and recall using benchmark dataset. The results show that by augmenting knowledge-based with classical model we can improve annotation accuracy by removing irrelevant keywords.}, address = {New York, NY, USA}, author = {Jin, Yohan and Khan, Latifur and Wang, Lei and Awad, Mamoun}, booktitle = {MULTIMEDIA '05: Proceedings of the 13th annual ACM international conference on Multimedia}, citeulike-article-id = {4010908}, citeulike-linkout-0 = {http://portal.acm.org/citation.cfm?id=1101149.1101305}, citeulike-linkout-1 = {http://dx.doi.org/10.1145/1101149.1101305}, doi = {10.1145/1101149.1101305}, isbn = {1-59593-044-2}, keywords = {annotation, cbir, external\_kb, wordnet}, location = {Hilton, Singapore}, pages = {706--715}, posted-at = {2009-02-05 10:59:30}, priority = {2}, publisher = {ACM}, title = {Image annotations by combining multiple evidence \& wordNet}, url = {http://dx.doi.org/10.1145/1101149.1101305}, year = {2005} }

TY - CONF ID - citeulike:4010908 L3 - citeulike-article-id:4010908 N2 - The development of technology generates huge amounts of non-textual information, such as images. An efficient image annotation and retrieval system is highly desired. Clustering algorithms make it possible to represent visual features of images with finite symbols. Based on this, many statistical models, which analyze correspondence between visual features and words and discover hidden semantics, have been published. These models improve the annotation and retrieval of large image databases. However, current state of the art including our previous work produces too many irrelevant keywords for images during annotation. In this paper, we propose a novel approach that augments the classical model with generic knowledge-based, WordNet. Our novel approach strives to prune irrelevant keywords by the usage of WordNet. To identify irrelevant keywords, we investigate various semantic similarity measures between keywords and finally fuse outcomes of all these measures together to make a final decision using Dempster-Shafer evidence combination. We have implemented various models to link visual tokens with keywords based on knowledge-based, WordNet and evaluated performance using precision, and recall using benchmark dataset. The results show that by augmenting knowledge-based with classical model we can improve annotation accuracy by removing irrelevant keywords. CY - Hilton, Singapore SN - 1-59593-044-2 AD - New York, NY, USA EP - 715 TI - Image annotations by combining multiple evidence \& wordNet PB - ACM T2 - MULTIMEDIA '05: Proceedings of the 13th annual ACM international conference on Multimedia SP - 706 KW - annotation KW - cbir KW - external_kb KW - wordnet AU - Jin, Yohan AU - Khan, Latifur AU - Wang, Lei AU - Awad, Mamoun PY - 2005/// UR - http://dx.doi.org/10.1145/1101149.1101305 ER -