Self-supervised Monocular Depth Estimation Robust to Reflective Surface Leveraged by Triplet Mining

TL;DR

This paper introduces a novel reflection-aware training strategy for self-supervised monocular depth estimation that improves accuracy on reflective surfaces by using triplet mining and knowledge distillation.

Contribution

It proposes a triplet mining loss and knowledge distillation method specifically designed to handle reflective surfaces in self-supervised depth estimation.

Findings

Enhanced depth accuracy on reflective surfaces.

Outperforms state-of-the-art baselines.

Robust depth estimation across various datasets.

Abstract

Self-supervised monocular depth estimation (SSMDE) aims to predict the dense depth map of a monocular image, by learning depth from RGB image sequences, eliminating the need for ground-truth depth labels. Although this approach simplifies data acquisition compared to supervised methods, it struggles with reflective surfaces, as they violate the assumptions of Lambertian reflectance, leading to inaccurate training on such surfaces. To tackle this problem, we propose a novel training strategy for an SSMDE by leveraging triplet mining to pinpoint reflective regions at the pixel level, guided by the camera geometry between different viewpoints. The proposed reflection-aware triplet mining loss specifically penalizes the inappropriate photometric error minimization on the localized reflective regions while preserving depth accuracy in non-reflective areas. We also incorporate a reflection-aware knowledge distillation method that enables a student model to selectively learn the pixel-level knowledge from reflective and non-reflective regions. This results in robust depth estimation across areas. Evaluation results on multiple datasets demonstrate that our method effectively enhances depth quality on reflective surfaces and outperforms state-of-the-art SSMDE baselines.