Unsupervisedly Training GANs for Segmenting Digital Pathology with Automatically Generated Annotations
This addresses the need for automated segmentation in renal pathology without manual annotations, but it is incremental as it builds on existing GAN-based semi-supervised methods.
The paper tackles the problem of fully unsupervised segmentation in digital pathology by generating simulated label data and using adversarial models for image-to-image translation, with experiments on glomeruli segmentation providing proof of concept and showing that the label data creation strategy affects GAN training stability.
Recently, generative adversarial networks exhibited excellent performances in semi-supervised image analysis scenarios. In this paper, we go even further by proposing a fully unsupervised approach for segmentation applications with prior knowledge of the objects' shapes. We propose and investigate different strategies to generate simulated label data and perform image-to-image translation between the image and the label domain using an adversarial model. Specifically, we assess the impact of the annotation model's accuracy as well as the effect of simulating additional low-level image features. For experimental evaluation, we consider the segmentation of the glomeruli, an application scenario from renal pathology. Experiments provide proof of concept and also confirm that the strategy for creating the simulated label data is of particular relevance considering the stability of GAN trainings.