Symbolic Graph Inference for Compound Scene Understanding

TL;DR

This work proposes a novel approach that reasons over a scene's scene- and knowledge-graph, capturing spatial information while being able to utilize general domain knowledge in a joint graph search.

Abstract

Scene understanding is a fundamental capability needed in many domains, ranging from question-answering to robotics. Unlike recent end-to-end approaches that must explicitly learn varying compositions of the same scene, our method reasons over their constituent objects and analyzes their arrangement to infer a scene's meaning. We propose a novel approach that reasons over a scene's scene- and knowledge-graph, capturing spatial information while being able to utilize general domain knowledge in a joint graph search. Empirically, we demonstrate the feasibility of our method on the ADE20K dataset and compare it to current scene understanding approaches.

Figures (2)

• Figure 1: The overview of our multi-graph reasoning approach which uses both the scene and knowledge graph and jointly reasons over them.
• Figure 2: Qualitative examples