Glomeruli Segmentation
Overview:
The Kidney Pathology Image segmentation (KPIs) challenge seeks to identify the state-of-the-art segmentation method for glomeruli identification in various Chronic kidney diseases (CKD). Participants are tasked with developing algorithms capable of precisely segmenting glomeruli at a pixel level, adapting to different CKD disease models and tissue conditions. This includes the ability to distinguish the glomeruli from surrounding tissue components under diverse preparation scenarios, thus demonstrating the versatility and precision of their approach. Additionally, the competition seeks solutions that are innovative in handling potential challenges such as variations in glomeruli size, shape, and structural integrity due to disease states or preparation techniques.
The challenge comprises two tasks aimed at advancing the field of kidney pathology image analysis:
Patch-level Segmentation: Participants will focus on the segmentation of glomeruli within specific image patches.
Whole Slide Image Segmentation: Participants are required to segment glomeruli across entire kidney slide images.
Participants have the flexibility to engage in both tasks or may choose to concentrate on a single task that aligns with their specific area of expertise or interest. The task examples are shown below,
Task 1 Patch-level Diseased Glomeruli Segmentation
Task 2 WSI-level Diseased Glomeruli Segmentation
Data:
Validation and testing data will be released for building and evaluating segmentation algorithms.
* Note that you must be part of the KPIs 2024 Data Access Team in order to view and download the KPIs 2024 datasets. To request data access, please fill in the form.