Leveraging semantic segmentation for hybrid image retrieval methods

Content-based image retrieval (CBIR) is the task of finding images in a database that are most similar to input query based on its visual characteristics. Several methods from state art (bag words, VLAD, etc.) or recent deep leaning try solve CBIR problem. In particular, learning new field and used for several vision applications including CBIR. But, even with increase performance algorithms, this problem still challenge computer vision. work, we propose three different methodologies combining learning-based semantic segmentation features. We show experimentally by exploiting information context leads an accuracy. study proposed approach eight datasets (Wang, Corel-10k, Corel-5k, GHIM-10K, MSRC V1, V2, Linnaeus NUS-WIDE)