Glo-In-One: Holistic Glomerular Detection, Segmentation, and Lesion Characterization with Large-scale Web Image Mining

TL;DR

This paper presents Glo-In-One, a comprehensive toolkit for automatic detection, segmentation, and lesion characterization of glomeruli in digital kidney pathology images, enhanced by large-scale web image mining and self-supervised learning.

Contribution

The paper introduces a user-friendly toolkit for holistic glomerular analysis and a large unlabeled image dataset to improve deep learning models with minimal annotated data.

Findings

Detection precision of 0.627 with circle representations

Segmentation Dice coefficient of 0.955

Effective self-supervised learning with only 10% labeled data

Abstract

The quantitative detection, segmentation, and characterization of glomeruli from high-resolution whole slide imaging (WSI) play essential roles in the computer-assisted diagnosis and scientific research in digital renal pathology. Historically, such comprehensive quantification requires extensive programming skills in order to be able to handle heterogeneous and customized computational tools. To bridge the gap of performing glomerular quantification for non-technical users, we develop the Glo-In-One toolkit to achieve holistic glomerular detection, segmentation, and characterization via a single line of command. Additionally, we release a large-scale collection of 30,000 unlabeled glomerular images to further facilitate the algorithmic development of self-supervised deep learning. The inputs of the Glo-In-One toolkit are WSIs, while the outputs are (1) WSI-level multi-class circle glomerular detection results (which can be directly manipulated with ImageScope), (2) glomerular image patches with segmentation masks, and (3) different lesion types. To leverage the performance of the Glo-In-One toolkit, we introduce self-supervised deep learning to glomerular quantification via large-scale web image mining. The GGS fine-grained classification model achieved a decent performance compared with baseline supervised methods while only using 10% of the annotated data. The glomerular detection achieved an average precision of 0.627 with circle representations, while the glomerular segmentation achieved a 0.955 patch-wise Dice Similarity Coefficient (DSC).