GasMono: Geometry-Aided Self-Supervised Monocular Depth Estimation for Indoor Scenes

TL;DR

GasMono introduces a novel framework combining geometry-based pose refinement and vision transformer-based depth guidance to improve self-supervised monocular depth estimation in indoor scenes, achieving state-of-the-art results.

Contribution

The paper proposes a new method that refines coarse camera poses during training and integrates vision transformers with self-distillation to enhance depth estimation accuracy.

Findings

Outperforms previous methods on multiple indoor datasets.

Achieves state-of-the-art accuracy in monocular depth estimation.

Demonstrates strong generalization across diverse indoor scenes.

Abstract

This paper tackles the challenges of self-supervised monocular depth estimation in indoor scenes caused by large rotation between frames and low texture. We ease the learning process by obtaining coarse camera poses from monocular sequences through multi-view geometry to deal with the former. However, we found that limited by the scale ambiguity across different scenes in the training dataset, a na\"ive introduction of geometric coarse poses cannot play a positive role in performance improvement, which is counter-intuitive. To address this problem, we propose to refine those poses during training through rotation and translation/scale optimization. To soften the effect of the low texture, we combine the global reasoning of vision transformers with an overfitting-aware, iterative self-distillation mechanism, providing more accurate depth guidance coming from the network itself. Experiments on NYUv2, ScanNet, 7scenes, and KITTI datasets support the effectiveness of each component in our framework, which sets a new state-of-the-art for indoor self-supervised monocular depth estimation, as well as outstanding generalization ability. Code and models are available at https://github.com/zxcqlf/GasMono