A Critical Synthesis of Uncertainty Quantification and Foundation Models in Monocular Depth Estimation

S. Landgraf 1,, R. Qin 2, M. Ulrich 1 · 2025

Read A Critical Synthesis of Uncertainty Quantification and Foundation Models in Monocular Depth Estimation by S. Landgraf 1,, R. Qin 2, M. Ulrich 1 (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-08188.txt.

Description

Steven Landgraf and co-authors fuse uncertainty tools with a top depth model in a 2025 synthesis paper. Guessing distance from a single photo has leapt forward with foundation models, yet even the best still make grave errors on absolute distance. Knowing when a model is unsure is key to safe use in cars and robots.

Five uncertainty methods were paired with the DepthAnythingV2 model and tested across four varied datasets. Fine-tuning with a Gaussian negative log-likelihood loss stood out as the most promising route. It gave trustworthy uncertainty readings while holding accuracy steady, at training and run costs on par with the plain model. The authors frame the study as groundwork for explainable machine vision, with clear next targets in segmentation and pose estimation.

Catalog record

CREATORS. Landgraf 1,, R. Qin 2, M. Ulrich 1
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE52 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-08188

Browse the full library · About the Tower of Babel