↓Skip to main content
Fernando Quintero

Enhancing Corneal Endothelium Segmentation: A Comparison of Classification and Distance-Map Regression UNets

· loading · loading · ·
Fernando Quintero👨‍💻
, 
Juan Sierra
, 
Kevin Mendoza
, 
Lenny A. Romero
, 
Andres G. Marrugo
Publication Presentation
Fernando Quintero
Author
Fernando Quintero
Researcher / Computer Vision / ML

IEEE Colombian Conference on Communications and Computing (COLCOM), 2023

🔗Paper

Abstract:
#

This paper compares two UNet-based architectures for corneal endothelium segmentation: a classification approach (UNet-mask) and a distance-map regression approach (UNet-dm). Our results show that the UNet-dm outperforms the UNet-mask with an average Dice coefficient of 0.8180 compared to 0.6583. Moreover, the UNet-dm model generates well-defined cell boundaries and produces mor-phometric parameters closer to the reference values. This study highlights the potential of distance-map regression-based UNet models for accurate corneal endothelium segmentation.

(a) CE image acquired by specular microscopy, (b) Ground truth segmentation, (c) Watersheed segmentation (UNet-mask).
(a) CE image acquired by specular microscopy, (b) Ground truth segmentation, (c) Watersheed segmentation (UNet-mask).

Citation:
#

F. J. Quintero, J. Sierra, K. D. Mendoza, L. A. Romero and A. G. Marrugo, “Enhancing Corneal Endothelium Segmentation: A Comparison of Classification and Distance-Map Regression UNets,” 2023 IEEE Colombian Conference on Communications and Computing (COLCOM), Bogota, Colombia, 2023, pp. 1-6, doi: 10.1109/COLCOM59909.2023.10334249.