Automatic Generation of Synthetic Colonoscopy Videos for Domain Randomization

TL;DR

This paper introduces a method to generate synthetic colonoscopy videos with diverse appearances and anatomical variations, aiming to improve machine learning models' generalization for guidance systems.

Contribution

It presents a novel approach for synthesizing colonoscopy videos that incorporate domain randomization to enhance training data diversity.

Findings

Synthetic videos improve model robustness to variations.

Enhanced generalization in colonoscopy guidance systems.

Synthetic data mimics real-world imaging conditions.

Abstract

An increasing number of colonoscopic guidance and assistance systems rely on machine learning algorithms which require a large amount of high-quality training data. In order to ensure high performance, the latter has to resemble a substantial portion of possible configurations. This particularly addresses varying anatomy, mucosa appearance and image sensor characteristics which are likely deteriorated by motion blur and inadequate illumination. The limited amount of readily available training data hampers to account for all of these possible configurations which results in reduced generalization capabilities of machine learning models. We propose an exemplary solution for synthesizing colonoscopy videos with substantial appearance and anatomical variations which enables to learn discriminative domain-randomized representations of the interior colon while mimicking real-world settings.