MICC short paper “Image Retrieval Using Multi-Scale CNN Feature Pooling” received the best poster award at ICMR 2020.
In this work we have presented an end-to-end trainable network architecture that exploits a novel multi-scale local pooling based on NetVLAD and a triplet mining procedure based on samples difficulty to obtain an effective image representation. Extensive experiments show that our approach is able to reach state-of-the-art results on three standard datasets.
The ACM International Conference on Multimedia Retrieval (ICMR) was created in 2011 from the merger of the ACM CIVR (International Conference on Image and Video Retrieval) and ACM MIR (International Conference on Multimedia Information Retrieval). ACM ICMR serves to illuminate the state of the art in multimedia (e.g., text, image, video, audio, sensor data, 3D) retrieval.
Link to paper: https://dl.acm.org/doi/
Bibliography: Federico Vaccaro, Marco Bertini, Tiberio Uricchio, and Alberto Del Bimbo. 2020. Image Retrieval using Multi-scale CNN Features Pooling. In Proceedings of the 2020 International Conference on Multimedia Retrieval (ICMR ’20). Association for Computing Machinery, New York, NY, USA, 311–315. DOI:https://doi.org/10.1145/