Instance-aware Exploration-Verification-Exploitation for Instance ImageGoal Navigation

TL;DR

This paper introduces a novel Instance-aware Exploration-Verification-Exploitation framework for instance image goal navigation, enabling agents to better identify and reach specific objects in complex environments, outperforming previous methods.

Contribution

The paper proposes a new modular navigation framework that actively switches among exploration, verification, and exploitation to improve instance-level navigation performance.

Findings

Outperforms previous state-of-the-art on HM3D-SEM dataset

Achieves higher success rates with classical and robust segmentation models

Demonstrates effective instance identification during navigation

Abstract

As a new embodied vision task, Instance ImageGoal Navigation (IIN) aims to navigate to a specified object depicted by a goal image in an unexplored environment. The main challenge of this task lies in identifying the target object from different viewpoints while rejecting similar distractors. Existing ImageGoal Navigation methods usually adopt the simple Exploration-Exploitation framework and ignore the identification of specific instance during navigation. In this work, we propose to imitate the human behaviour of ``getting closer to confirm" when distinguishing objects from a distance. Specifically, we design a new modular navigation framework named Instance-aware Exploration-Verification-Exploitation (IEVE) for instance-level image goal navigation. Our method allows for active switching among the exploration, verification, and exploitation actions, thereby facilitating the agent in making reasonable decisions under different situations. On the challenging HabitatMatterport 3D semantic (HM3D-SEM) dataset, our method surpasses previous state-of-the-art work, with a classical segmentation model (0.684 vs. 0.561 success) or a robust model (0.702 vs. 0.561 success)