Understanding Spatial Relations through Multiple Modalities

TL;DR

This paper introduces a multimodal approach to infer explicit and implicit spatial relations between objects in images, combining visual and textual data to improve understanding and reasoning in spatial contexts.

Contribution

It presents a novel model that integrates visual and textual modalities to predict spatial relations, enhancing accuracy and generalization over prior models.

Findings

Improved prediction accuracy over language-only models.

Enhanced ability to handle unseen objects and relations.

Demonstrated effectiveness in diverse spatial reasoning tasks.

Abstract

Recognizing spatial relations and reasoning about them is essential in multiple applications including navigation, direction giving and human-computer interaction in general. Spatial relations between objects can either be explicit -- expressed as spatial prepositions, or implicit -- expressed by spatial verbs such as moving, walking, shifting, etc. Both these, but implicit relations in particular, require significant common sense understanding. In this paper, we introduce the task of inferring implicit and explicit spatial relations between two entities in an image. We design a model that uses both textual and visual information to predict the spatial relations, making use of both positional and size information of objects and image embeddings. We contrast our spatial model with powerful language models and show how our modeling complements the power of these, improving prediction accuracy and coverage and facilitates dealing with unseen subjects, objects and relations.