Unsupervised object discovery for instance recognition

TL;DR

This paper introduces an unsupervised salient region detection method that enhances instance recognition by focusing on discriminative and common patterns, improving retrieval accuracy in cluttered large-scale datasets.

Contribution

It presents a novel unsupervised salient region detection technique based on graph centrality of CNN features, improving object retrieval performance.

Findings

Enhanced retrieval accuracy in cluttered datasets

Better focus on discriminative regions improves instance recognition

Effective in large collections with small objects

Abstract

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 region 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.