Author: Tomasz Strzoda Published At: 25.05.2026 Updated At: 27.05.2026
Seminar: Julia Merta
On May 20, during a department seminar, Julia Merta, M.Sc., presented HistoKit, a tool designed for processing whole-slide microscope (WSI) images. The software enables the detection of tissue areas and artifacts using the GrandQC deep learning model, as well as the normalization of staining and the division of images into smaller segments to prepare data for training deep learning models.
The presentation also included a comparison of the effectiveness of the original version of the GrandQC algorithm, the HistoQC tool, and HistoKit in terms of artifact detection. The analysis was conducted, among other things, on a self-labeled dataset comprising histopathological images from open TCGA databases and images from the Compass NMD project.
