Initial Results Refinement Based on Contextual Information for Interactive Image Retrieval Systems

• Luciano Araujo Dourado Filho UEFS
• Rodrigo Tripodi Calumby UEFS

DOI: https://doi.org/10.5753/sbbd.2019.8837

Abstract

This paper proposes the exploitation of contextual information to refine the initial set of results in interactive image retrieval systems. This context based method differs from traditional approaches given it considers the rich information present in the relationships between multiple images, rather than simply using traditional methods that computes pairwise distances between two images. In this work, an experimental evaluation of the effectiveness of this technique in different Web search scenarios was carried out. The results showed that the evaluated method was able to significantly improve the effectiveness of the retrieval, specially when applied in conjunction with similarity measures of greater distinctive power.

Keywords: Contextual Information, Information Retrieval, Cold Start

References

Arni, T., Clough, P., Sanderson, M., and Grubinger, M. (2009). Overview of the imageclefphoto 2008 photographic retrieval task. In Evaluating Systems for Multilingual and Multimodal Information Access, pages 500–511. Springer Berlin Heidelberg. DOI: https:/doi.org/10.1145/2964797.2964811

Baeza-Yates, R. and Ribeiro-Neto, B. (2011). Modern Information Retrieval: The Concepts and Technology Behind Search. Addison-Wesley, USA, 2nd edition. DOI: https:/doi.org/10.5860/choice.48-6950

Calumby, R. T., da S. Torres, R., and Gonçalves, M. A. (2014). Multimodal retrieval with relevance feedback based on genetic programming. Multimedia Tools Appl, (69):991–1019. DOI: https:/doi.org/10.1007/s11042-012-1152-7

Calumby, R. T., da S. Torres, R., and Gonçalves, M. A. (2017). Diversity-based interactive learning meets multimodality. Neurocomputing, (259):159–175. DOI: https:/doi.org/10.13039/501100002322

Lewis, J., Ossowski, S., Hicks, J., Errami, M., and Garner, H. R. (2006). Text similarity: an alternative way to search MEDLINE. Bioinformatics, 22(18):2298–2304. DOI: https:/doi.org/10.1093/bioinformatics/btl388

Pedronette, D. C. G., da S. Torres, R., and Calumby, R. T. (2014). Using contextual spaces for image re-ranking and rank aggregation. Multimedia Tools Appl, 69(3):689–716. DOI: https:/doi.org/10.1007/s11042-012-1115-z

Santos, K., de Almeida, H. M., Gonc¸alves, M. A., and da Silva Torres, R. (2009). Recuperação de imagens da web utilizando múltiplas evidências textuais e programação genética. In XXIV SBBD, pages 91–105. DOI: https:/doi.org/

Torres, R. and Falcão, A. X. (2006). Content-based image retrieval: Theory and applications. Revista de Informática Teórica e Aplicada, (161-185):13(2). DOI: https:/doi.org/