Detecting Visual Relationships with Deep Relational Networks

TLDR

Object relationships are essential for image understanding, yet existing deep learning methods struggle to reason about them due to the high diversity of visual appearances and the large number of distinct relationship categories. The authors propose an integrated framework to address the challenge of reasoning about object relationships. The framework centers on a Deep Relational Network that exploits statistical dependencies between objects and their relationships. On two large datasets, the method substantially outperforms state‑of‑the‑art approaches.

Abstract

Relationships among objects play a crucial role in image understanding. Despite the great success of deep learning techniques in recognizing individual objects, reasoning about the relationships among objects remains a challenging task. Previous methods often treat this as a classification problem, considering each type of relationship (e.g. ride) or each distinct visual phrase (e.g. person-ride-horse) as a category. Such approaches are faced with significant difficulties caused by the high diversity of visual appearance for each kind of relationships or the large number of distinct visual phrases. We propose an integrated framework to tackle this problem. At the heart of this framework is the Deep Relational Network, a novel formulation designed specifically for exploiting the statistical dependencies between objects and their relationships. On two large data sets, the proposed method achieves substantial improvement over state-of-the-art.