REGNav: Room Expert Guided Image-Goal Navigation

TL;DR

This work addresses image-goal navigation when goal and current observations may reside in different rooms by introducing a Room Expert that predicts whether two images come from the same room using unsupervised room-style representations. The method, REGNav, follows a two-stage learning paradigm: offline pre-training of the Room Expert on unlabeled indoor imagery and subsequent navigation policy learning with the expert frozen, employing either implicit or explicit fusion to inject room-relational information into decision-making. Empirical results on Gibson show REGNav achieving state-of-the-art SPL and SR, with explicit fusion providing the strongest performance, while cross-domain tests on MP3D and HM3D demonstrate robust generalization without additional finetuning. The approach highlights the value of spatial priors derived from room styles to guide RGB-only navigation, offering a scalable path toward more reliable embodied agents in unseen environments.

Abstract

Image-goal navigation aims to steer an agent towards the goal location specified by an image. Most prior methods tackle this task by learning a navigation policy, which extracts visual features of goal and observation images, compares their similarity and predicts actions. However, if the agent is in a different room from the goal image, it's extremely challenging to identify their similarity and infer the likely goal location, which may result in the agent wandering around. Intuitively, when humans carry out this task, they may roughly compare the current observation with the goal image, having an approximate concept of whether they are in the same room before executing the actions. Inspired by this intuition, we try to imitate human behaviour and propose a Room Expert Guided Image-Goal Navigation model (REGNav) to equip the agent with the ability to analyze whether goal and observation images are taken in the same room. Specifically, we first pre-train a room expert with an unsupervised learning technique on the self-collected unlabelled room images. The expert can extract the hidden room style information of goal and observation images and predict their relationship about whether they belong to the same room. In addition, two different fusion approaches are explored to efficiently guide the agent navigation with the room relation knowledge. Extensive experiments show that our REGNav surpasses prior state-of-the-art works on three popular benchmarks.

Figures (3)

• Figure 1: We solve the task of image-goal navigation, where an agent (the yellow robot) is required to navigate to a location depicted by a goal image. To accomplish this, our agent tries to compare the current observation with the goal image and tease out whether the current location is in the same room with the goal image before executing actions.
• Figure 2: The overview of our REGNav. (a) Pre-training the Room Expert offline. We employ an unsupervised clustering method to train a style encoder and a relation network to extract style representation and predict the relationships. We use the constraints set deduced from the unlabeled data to refine the feature distance matrix to obtain more reliable pseudo labels. (b) The image-goal navigation architecture with Room Expert. We lock the Room Expert and proceed to train the visual encoder and navigation policy. The visual feature extractor regards the channel concatenation of the observation and goal image as input. The navigation policy takes the concatenation of the relation flag (2-dimension) and the fused feature as input.
• Figure 3: The visualization results of example episodes from a top-down view. The lines originating from the green locations refer to the agent's trajectories, where the colour changes as the steps. The grey regions on the top-down map represent the areas explored by the agent's camera. Compared with the baseline, our REGNav plans more efficient navigation paths.