Weakly-supervised learning of visual relations

TL;DR

This paper presents a weakly-supervised approach for learning visual object relations using discriminative clustering, novel features, and a new dataset, achieving state-of-the-art results especially in zero-shot relation recognition.

Contribution

It introduces a weakly-supervised discriminative clustering model, new visual features for relation modeling, and a challenging dataset for evaluating visual relation retrieval.

Findings

State-of-the-art performance on visual relation dataset

Significant improvement in zero-shot relation recognition

Effective relation modeling with weak supervision

Abstract

This paper introduces a novel approach for modeling visual relations between pairs of objects. We call relation a triplet of the form (subject, predicate, object) where the predicate is typically a preposition (eg. 'under', 'in front of') or a verb ('hold', 'ride') that links a pair of objects (subject, object). Learning such relations is challenging as the objects have different spatial configurations and appearances depending on the relation in which they occur. Another major challenge comes from the difficulty to get annotations, especially at box-level, for all possible triplets, which makes both learning and evaluation difficult. The contributions of this paper are threefold. First, we design strong yet flexible visual features that encode the appearance and spatial configuration for pairs of objects. Second, we propose a weakly-supervised discriminative clustering model to learn relations from image-level labels only. Third we introduce a new challenging dataset of unusual relations (UnRel) together with an exhaustive annotation, that enables accurate evaluation of visual relation retrieval. We show experimentally that our model results in state-of-the-art results on the visual relationship dataset significantly improving performance on previously unseen relations (zero-shot learning), and confirm this observation on our newly introduced UnRel dataset.