On the Uncertain Single-View Depths in Colonoscopies

TL;DR

This paper explores Bayesian deep networks for single-view depth estimation in colonoscopies, addressing domain challenges and proposing a teacher-student method that incorporates uncertainty for improved depth learning.

Contribution

It provides the first analysis of Bayesian networks for colonoscopy depth estimation and introduces a novel teacher-student approach considering teacher uncertainty.

Findings

Bayesian networks reveal domain adaptation challenges in colonoscopy depth estimation.

Self-supervised methods show promise over supervised approaches in certain datasets.

Teacher-student models improve depth accuracy by leveraging uncertainty information.

Abstract

Estimating depth information from endoscopic images is a prerequisite for a wide set of AI-assisted technologies, such as accurate localization and measurement of tumors, or identification of non-inspected areas. As the domain specificity of colonoscopies -- deformable low-texture environments with fluids, poor lighting conditions and abrupt sensor motions -- pose challenges to multi-view 3D reconstructions, single-view depth learning stands out as a promising line of research. Depth learning can be extended in a Bayesian setting, which enables continual learning, improves decision making and can be used to compute confidence intervals or quantify uncertainty for in-body measurements. In this paper, we explore for the first time Bayesian deep networks for single-view depth estimation in colonoscopies. Our specific contribution is two-fold: 1) an exhaustive analysis of scalable Bayesian networks for depth learning in different datasets, highlighting challenges and conclusions regarding synthetic-to-real domain changes and supervised vs. self-supervised methods; and 2) a novel teacher-student approach to deep depth learning that takes into account the teacher uncertainty.