GeoDiff: Geometry-Guided Diffusion for Metric Depth Estimation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Pham, Tuan, Le, Thanh-Tung, Xie, Xiaohui, Mandt, Stephan
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909861010210816
author Pham, Tuan
Le, Thanh-Tung
Xie, Xiaohui
Mandt, Stephan
author_facet Pham, Tuan
Le, Thanh-Tung
Xie, Xiaohui
Mandt, Stephan
contents We introduce a novel framework for metric depth estimation that enhances pretrained diffusion-based monocular depth estimation (DB-MDE) models with stereo vision guidance. While existing DB-MDE methods excel at predicting relative depth, estimating absolute metric depth remains challenging due to scale ambiguities in single-image scenarios. To address this, we reframe depth estimation as an inverse problem, leveraging pretrained latent diffusion models (LDMs) conditioned on RGB images, combined with stereo-based geometric constraints, to learn scale and shift for accurate depth recovery. Our training-free solution seamlessly integrates into existing DB-MDE frameworks and generalizes across indoor, outdoor, and complex environments. Extensive experiments demonstrate that our approach matches or surpasses state-of-the-art methods, particularly in challenging scenarios involving translucent and specular surfaces, all without requiring retraining.
format Preprint
id arxiv_https___arxiv_org_abs_2510_18291
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle GeoDiff: Geometry-Guided Diffusion for Metric Depth Estimation
Pham, Tuan
Le, Thanh-Tung
Xie, Xiaohui
Mandt, Stephan
Computer Vision and Pattern Recognition
We introduce a novel framework for metric depth estimation that enhances pretrained diffusion-based monocular depth estimation (DB-MDE) models with stereo vision guidance. While existing DB-MDE methods excel at predicting relative depth, estimating absolute metric depth remains challenging due to scale ambiguities in single-image scenarios. To address this, we reframe depth estimation as an inverse problem, leveraging pretrained latent diffusion models (LDMs) conditioned on RGB images, combined with stereo-based geometric constraints, to learn scale and shift for accurate depth recovery. Our training-free solution seamlessly integrates into existing DB-MDE frameworks and generalizes across indoor, outdoor, and complex environments. Extensive experiments demonstrate that our approach matches or surpasses state-of-the-art methods, particularly in challenging scenarios involving translucent and specular surfaces, all without requiring retraining.
title GeoDiff: Geometry-Guided Diffusion for Metric Depth Estimation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2510.18291