Unsupervised Selective Rank Fusion for Content-based Image Retrieval

  • Lucas Pascotti Valem UNESP
  • Daniel Carlos Guimarães Pedronette UNESP

Resumo


The CBIR (Content-Based Image Retrieval) systems are one of the main solutions for image retrieval tasks. These systems are mainly supported by the use of different visual features and machine learning methods. As distinct features produce complementary ranking results with different effectiveness performance, a promising solution consists in combining them. However, how to decide which visual features to combine is a very challenging task, especially when no training data is available. This work proposes three novel methods for selecting and combining ranked lists by estimating their effectiveness in an unsupervised way. The approaches were evaluated in five different image collections and several descriptors, achieving results comparable or superior to the state-of-the-art in most of the scenarios.

Palavras-chave: Content-based Image Retrieval, Unsupervised Learning, Re-ranking, Rank-aggregation, Feature Selection, Feature Fusion

Referências

Almeida, J., Valem, L. P., and Pedronette, D. C. G. (2017). A rank aggregation framework for video interestingness prediction. In Image Analysis and Processing-ICIAP 2017, pages 3-14. Springer International Publishing.

Pedronette, D. C. G., Valem, L. P., Almeida, J., and da Silva Torres, R. (2019). Multimedia retrieval through unsupervised hypergraph-based manifold ranking. IEEE Transactions on Image Processing, 28(12):5824-5838.

Valem, L. and Pedronette, D. (2019a). Unsupervised selective rank fusion on content-based image retrieval. In Anais Estendidos da XXXII Conference on Graphics, Patterns and Images, pages 63-69, Porto Alegre, RS, Brasil. SBC.

Valem, L. P., Oliveira, C. R. D., Pedronette, D. C. G., and Almeida, J. (2018). Unsupervised similarity learning through rank correlation and knn sets. In ACM Trans. Multimedia Comput. Commun. Appl., volume 14, pages 80:1-80:23, New York, NY, USA. ACM.

Valem, L. P. and Pedronette, D. C. G. (2017). Selection and combination of unsupervised learning methods for image retrieval. In Proceedings of the 15th International Workshop on Content-Based Multimedia Indexing, CBMI ’17, pages 27:1-27:6.

Valem, L. P. and Pedronette, D. C. G. (2019b). Combinação Seletiva Não Supervisionada de Listas Ranqueadas Aplicada a Busca de Imagens pelo Conteúdo. Dissertation (M.Sc. in Computer Science), UNESP (Universidade Estadual Paulista Julio de Mesquita Filho), Rio Claro, São Paulo, Brazil.

Valem, L. P. and Pedronette, D. C. G. (2019c). An unsupervised genetic algorithm framework for rank selection and fusion on image retrieval. In Proceedings of the 2019 on International Conference on Multimedia Retrieval, ICMR ’19, pages 58-62, New York, NY, USA. ACM.

Valem, L. P. and Pedronette, D. C. G. (2020a). Graph-based selective rank fusion for unsupervised image retrieval. Pattern Recognition Letters, 135:82-89.

Valem, L. P. and Pedronette, D. C. G. (2020b). Unsupervised selective rank fusion for image retrieval tasks. Neurocomputing, 377:182-199.

Valem, L. P., Pedronette, D. C. G., Breve, F., and Guilherme, I. R. (2018). Manifold correlation graph for semi-supervised learning. In 2018 International Joint Conference on Neural Networks (IJCNN), pages 1-7.
Publicado
30/06/2020
VALEM, Lucas Pascotti; PEDRONETTE, Daniel Carlos Guimarães. Unsupervised Selective Rank Fusion for Content-based Image Retrieval. In: CONCURSO DE TESES E DISSERTAÇÕES (CTD), 33. , 2020, Cuiabá. Anais [...]. Porto Alegre: Sociedade Brasileira de Computação, 2020 . p. 61-66. ISSN 2763-8820. DOI: https://doi.org/10.5753/ctd.2020.11370.