Contextual Similarity Learning for Image Retrieval and Classification: Applications in Person Re-Identification

• Lucas Pascotti Valem UNESP / USP
• Daniel Carlos Guimarães Pedronette UNESP

DOI: https://doi.org/10.5753/ctd.2025.8145

Resumo

The rapid expansion of image collections has driven the need for advanced machine learning and image retrieval applications. However, many existing methods depend on large labeled datasets for training, which are costly. To address this, various techniques have been developed, with a key challenge being the effective definition of image similarity. Most approaches still rely on pairwise similarity measures, overlooking valuable contextual information. This work enhances image retrieval and classification by leveraging contextual similarity, proposing seven novel methods applied to general-purpose and person reidentification (Re-ID) datasets. Extensive experiments show that the proposed methods perform comparably to or better than state-of-the-art approaches.

Referências

Antelmi, A., Cordasco, G., Polato, M., Scarano, V., Spagnuolo, C., and Yang, D. (2023). A survey on hypergraph representation learning. ACM Computing Surveys, 56(1).

Liao, S. and Shao, L. (2022). Graph sampling based deep metric learning for generalizable person re-identification. In 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 7349–7358.

Peng, L., Qiang, B., and Wu, J. (2022). A survey: Image classification models based on convolutional neural networks. In 2022 14th International Conference on Computer Research and Development (ICCRD), pages 291–298.

Rodrigues, J. and Carbonera, J. (2024). Graph convolutional networks for image classification: Comparing approaches for building graphs from images. In 26th Int. Conference on Enterprise Information Systems - Vol. 1: ICEIS, pages 437–446. SciTePress.

Song, Z., Yang, X., Xu, Z., and King, I. (2023). Graph-based semi-supervised learning: A comprehensive review. IEEE Transactions on Neural Networks and Learning Systems, 34(11):8174–8194.

Srivastava, D., Singh, S. S., Rajitha, B., Verma, M., Kaur, M., and Lee, H.-N. (2023). Content-based image retrieval: A survey on local and global features selection, extraction, representation, and evaluation parameters. IEEE Access, 11:95410–95431.

Tripathi, S. and King, C. R. (2024). Contrastive learning: Big data foundations and applications. In Proceedings of the CODS-COMAD ’24, page 493–497, New York, NY, USA. Association for Computing Machinery.

Uelwer, T., Robine, J., Wagner, S. S., Höftmann, M., Upschulte, E., Konietzny, S., Behrendt, M., and Harmeling, S. (2023). A survey on self-supervised representation learning. arXiv preprint arXiv:2308.11455.

Valem, L. P., Kawai, V. A. S., Pereira-Ferrero, V. H., and Pedronette, D. C. G. (2022). A novel rank correlation measure for manifold learning on image retrieval and person re-id. In IEEE International Conference on Image Processing (ICIP), pages 1371–1375.

Valem, L. P., Pedronette, D., and Allili, M. S. (2024). Contrastive loss based on contextual similarity for image classification. In 19th International Symposium on Visual Computing (ISVC).

Valem, L. P. and Pedronette, D. C. G. (2021). A denoising convolutional neural network for self-supervised rank effectiveness estimation on image retrieval. In Proc. of the International Conference on Multimedia Retrieval, ICMR ’21, page 294–302. ACM.

Valem, L. P. and Pedronette, D. C. G. (2022). Person re-id through unsupervised hypergraph rank selection and fusion. Image Vision Computing, 123(C).

Valem, L. P., Pedronette, D. C. G., and Latecki, L. J. (2023a). Graph convolutional networks based on manifold learning for semi-supervised image classification. Computer Vision and Image Understanding, 227:103618.

Valem, L. P., Pedronette, D. C. G., and Latecki, L. J. (2023b). Rank flow embedding for unsupervised and semi-supervised manifold learning. IEEE Transactions on Image Processing, 32:2811–2826.

Valem, L. P., Pereira-Ferrero, V., and Pedronette, D. (2023c). Self-supervised regression for query performance prediction on image retrieval. In 2023 IEEE Sixth International Conference on Artificial Intelligence and Knowledge Engineering (AIKE), pages 95–98, Los Alamitos, CA, USA. IEEE Computer Society.