Impact of Balancing and Regularization on the Semantic Segmentation of Gleason Patterns

Eduardo Paraíso, Alexei Machado, Alexei Machado

2025

Abstract

This study investigates the impact of class balancing and regularization on improving the diagnostic agreement in prostate histological images. The U-Net models applied to the Prostate Cancer Grade Assessment dataset reveal that class balancing combined with traditional loss functions contributes to an increase of up to 6 percentage points in image agreement. Combining balancing and Focal Loss can increase image classification agreement by an average of 13 percentage points compared to using an imbalanced dataset with traditional loss functions. Notably, distinguishing between Gleason patterns 3 and 4 in medical image analysis is crucial, as this distinction not only directly influences clinical decisions and the prognosis of prostate cancer patients but also emphasizes the need for careful interpretation of the data.

Download


Paper Citation


in Harvard Style

Paraíso E. and Machado A. (2025). Impact of Balancing and Regularization on the Semantic Segmentation of Gleason Patterns. In Proceedings of the 18th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 2: HEALTHINF; ISBN 978-989-758-731-3, SciTePress, pages 382-389. DOI: 10.5220/0013105500003911


in Bibtex Style

@conference{healthinf25,
author={Eduardo Paraíso and Alexei Machado},
title={Impact of Balancing and Regularization on the Semantic Segmentation of Gleason Patterns},
booktitle={Proceedings of the 18th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 2: HEALTHINF},
year={2025},
pages={382-389},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013105500003911},
isbn={978-989-758-731-3},
}


in EndNote Style

TY - CONF

JO - Proceedings of the 18th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 2: HEALTHINF
TI - Impact of Balancing and Regularization on the Semantic Segmentation of Gleason Patterns
SN - 978-989-758-731-3
AU - Paraíso E.
AU - Machado A.
PY - 2025
SP - 382
EP - 389
DO - 10.5220/0013105500003911
PB - SciTePress