SIRI: Spatial Relation Induced Network For Spatial Description Resolution

TL;DR

This paper introduces SIRI, a novel network that models spatial relationships explicitly for language-guided localization in panoramic views, significantly improving accuracy over previous methods.

Contribution

The paper proposes a new spatial relationship induced network that mimics human spatial reasoning, incorporating object-level correlation, spatial relationship distillation, and global position priors.

Findings

Achieves 24% better accuracy than state-of-the-art on Touchdown dataset.

Effectively generalizes to an extended dataset with similar settings.

Improves spatial description resolution by explicit relationship modeling.

Abstract

Spatial Description Resolution, as a language-guided localization task, is proposed for target location in a panoramic street view, given corresponding language descriptions. Explicitly characterizing an object-level relationship while distilling spatial relationships are currently absent but crucial to this task. Mimicking humans, who sequentially traverse spatial relationship words and objects with a first-person view to locate their target, we propose a novel spatial relationship induced (SIRI) network. Specifically, visual features are firstly correlated at an implicit object-level in a projected latent space; then they are distilled by each spatial relationship word, resulting in each differently activated feature representing each spatial relationship. Further, we introduce global position priors to fix the absence of positional information, which may result in global positional reasoning ambiguities. Both the linguistic and visual features are concatenated to finalize the target localization. Experimental results on the Touchdown show that our method is around 24\% better than the state-of-the-art method in terms of accuracy, measured by an 80-pixel radius. Our method also generalizes well on our proposed extended dataset collected using the same settings as Touchdown.