Unsupervised Object Discovery for Instance Recognition

Oriane Simeoni, Ahmet Iscen, Giorgos Tolias, Yannis Avrithis, Ondrej Chum
2018 2018 IEEE Winter Conference on Applications of Computer Vision (WACV)  
Severe background clutter is challenging in many computer vision tasks, including large-scale image retrieval. Global descriptors, that are popular due to their memory and search efficiency, are especially prone to corruption by such a clutter. Eliminating the impact of the clutter on the image descriptor increases the chance of retrieving relevant images and prevents topic drift due to actually retrieving the clutter in the case of query expansion. In this work, we propose a novel salient
more » ... n detection method. It captures, in an unsupervised manner, patterns that are both discriminative and common in the dataset. Saliency is based on a centrality measure of a nearest neighbor graph constructed from regional CNN representations of dataset images. The descriptors derived from the salient regions improve particular object retrieval, most noticeably in a large collections containing small objects.
doi:10.1109/wacv.2018.00194 dblp:conf/wacv/SimeoniITAC18 fatcat:yqdibyg5tbdxnepmeuamtq2y6e