Related Experiment Video
Updated: Jan 9, 2026
09:16
Supervised Machine Learning for Semi-Quantification of Extracellular DNA in Glomerulonephritis
Published on: June 18, 2020
7.3K
A General Pipeline for Glomerulus Whole-Slide Image Segmentation
Summary
A new pipeline improves whole-slide image (WSI) glomerulus segmentation for kidney disease diagnosis. This method enhances detection accuracy, setting a new benchmark for WSI analysis.
Area of Science:
- Digital pathology
- Medical image analysis
- Nephrology
Background:
- Accurate glomerulus segmentation in whole-slide images (WSIs) is critical for diagnosing kidney diseases.
- Existing segmentation methods face challenges with glomeruli near patch borders in WSIs.
Purpose of the Study:
- To develop a general and practical pipeline to enhance glomerulus segmentation in WSIs.
- To improve both patch-level and WSI-level segmentation performance.
Main Methods:
- Implemented a pipeline leveraging stitching on overlapping patches to increase detection coverage.
- Conducted comprehensive evaluations using diverse segmentation models on two large datasets.
- Utilized over 30,000 glomerulus annotations for training and validation.
Main Results:
- The proposed pipeline significantly outperformed the previous state-of-the-art method.
- Achieved superior glomerulus segmentation results across both evaluated datasets.
- Established a new benchmark for glomerulus segmentation in WSIs.
Conclusions:
- The developed pipeline offers a robust solution for accurate glomerulus segmentation in digital pathology.
- This advancement has the potential to improve the diagnostic accuracy of kidney diseases through automated WSI analysis.
- The pipeline enhances segmentation by effectively handling glomeruli at patch image borders.
