IMPROVING TAG BASED IMAGE SEARCH WITH SOCIAL RE-RANKING STRATEGIES

     In this paper, a social re-ranking method for tag-based image retrieval is presented. The proposed system uses both inter-user re-ranking and intra-user re-ranking techniques to improve the quality of retrieved images. By introducing user information into the ranking process, the system is able to enhance the diversity of the search results. In addition, the number of views of social images is integrated into the regularization framework to improve the relevance of the retrieved results.      The experimental analysis shows that the proposed method is effective and also reduces the time required for retrieving images. However, during the inter-user ranking process, only the contribution of users is considered, while the similarity between users is not taken into account. Furthermore, several useful details available in the Flickr dataset, such as title information and time stamps, are not fully utilized. These aspects can be considered for further improvements in future work.

[1] X. Zhang, L. Zheng and Y. Yang, “Understanding Image Retrieval Re-Ranking: A Graph Neural Network Perspective,” arXiv preprint arXiv:2012.07620, 2020. [2] K. Mao, Z. Yang and J. Li, “Item Tagging for Information Retrieval: A Tripartite Graph Neural Network Approach,” arXiv preprint arXiv:2008.11567, 2020. [3] J. Li and S. Wang, “Deep Learning for Image Retrieval: A Survey,” IEEE Transactions on Neural Networks and Learning Systems, vol. 31, no. 10, pp. 3902–3920, 2020. [4] X. Qian, G. Liu and H. Wang, “Social Image Retrieval Based on Topic Diversity,” Multimedia Tools and Applications, vol. 80, pp. 25473–25492, 2021. [5] H. Liu, Y. Zhang and D. Xu, “Recent Developments of Content-Based Image Retrieval,” Neurocomputing, vol. 452, pp. 259–273, 2021. [6] S. Kumar and R. Patel, “Deep Learning Based Image Retrieval System Using Visual Features,” International Journal of Computer Applications, vol. 183, no. 45, pp. 20–25, 2021. [7] M. Brown and T. Wilson, “Semantic Image Retrieval Using Feature Fusion Techniques,” Journal of Visual Communication and Image Representation, vol. 75, 2021. [8] Y. Li, X. Zhao and H. Wang, “Graph Convolution Based Efficient Re-Ranking for Visual Retrieval,” IEEE Access, vol. 10, pp. 14562–14572, 2022. [9] A. Gupta and P. Singh, “Machine Learning Techniques for Image Retrieval Systems: A Survey,” Journal of Artificial Intelligence Research, vol. 74, pp. 415–436, 2022. [10] D. Roy and K. Das, “Improved Tag-Based Image Retrieval Using Semantic Analysis,” International Journal of Computer Science and Information Security, vol. 20, no. 3, pp. 55–62, 2022.