Diversity in image retrieval: DCU at ImageCLEFPhoto 2008
CLEF'08 Proceedings of the 9th Cross-language evaluation forum conference on Evaluating systems for multilingual and multimodal information access
| Hi-index | 0.00 |
Dublin City University (DCU) and University of Tampere (UTA) participated in ImageCLEF 2007 photographic retrieval task with several monolingual and bilingual runs. The approach was language independent with text retrieval utilizing fuzzy s-gram query translation and combined with visual retrieval. Data fusion was achieved through unsupervised query-time weight generation approaches. The baseline was a combination of dictionary-based query translation and visual retrieval, which achieved the best result. The best mixed modality runs using fuzzy s-gram translation reached on average around 83% of the baselines' performance. This approach was much closer at the early precision levels of P@10 and P@20. This suggests that our language independent approach could be a cheap alternative for cross-lingual image retrieval. Both sets of results further emphasize the merit in our query-time weight generation schemes for data fusion, with the fused runs exhibiting marked performance increases over single modalities without the use of prior training data.