ULS+: Data-driven Model Adaptation Enhances Lesion Segmentation

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Weber, Rianne, Rocholl, Niels, de Grauw, Max, Prokop, Mathias, Smit, Ewoud, Hering, Alessa
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917187144384512
author Weber, Rianne
Rocholl, Niels
de Grauw, Max
Prokop, Mathias
Smit, Ewoud
Hering, Alessa
author_facet Weber, Rianne
Rocholl, Niels
de Grauw, Max
Prokop, Mathias
Smit, Ewoud
Hering, Alessa
contents In this study, we present ULS+, an enhanced version of the Universal Lesion Segmentation (ULS) model. The original ULS model segments lesions across the whole body in CT scans given volumes of interest (VOIs) centered around a click-point. Since its release, several new public datasets have become available that can further improve model performance. ULS+ incorporates these additional datasets and uses smaller input image sizes, resulting in higher accuracy and faster inference. We compared ULS and ULS+ using the Dice score and robustness to click-point location on the ULS23 Challenge test data and a subset of the Longitudinal-CT dataset. In all comparisons, ULS+ significantly outperformed ULS. Additionally, ULS+ ranks first on the ULS23 Challenge test-phase leaderboard. By maintaining a cycle of data-driven updates and clinical validation, ULS+ establishes a foundation for robust and clinically relevant lesion segmentation models.
format Preprint
id arxiv_https___arxiv_org_abs_2601_02988
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle ULS+: Data-driven Model Adaptation Enhances Lesion Segmentation
Weber, Rianne
Rocholl, Niels
de Grauw, Max
Prokop, Mathias
Smit, Ewoud
Hering, Alessa
Computer Vision and Pattern Recognition
Artificial Intelligence
In this study, we present ULS+, an enhanced version of the Universal Lesion Segmentation (ULS) model. The original ULS model segments lesions across the whole body in CT scans given volumes of interest (VOIs) centered around a click-point. Since its release, several new public datasets have become available that can further improve model performance. ULS+ incorporates these additional datasets and uses smaller input image sizes, resulting in higher accuracy and faster inference. We compared ULS and ULS+ using the Dice score and robustness to click-point location on the ULS23 Challenge test data and a subset of the Longitudinal-CT dataset. In all comparisons, ULS+ significantly outperformed ULS. Additionally, ULS+ ranks first on the ULS23 Challenge test-phase leaderboard. By maintaining a cycle of data-driven updates and clinical validation, ULS+ establishes a foundation for robust and clinically relevant lesion segmentation models.
title ULS+: Data-driven Model Adaptation Enhances Lesion Segmentation
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2601.02988