Semi-Supervised Segmentation of Functional Tissue Units at the Cellular Level

TL;DR

This paper introduces a semi-supervised deep learning method for cellular-level tissue unit segmentation that effectively addresses domain differences and class imbalance, achieving state-of-the-art results.

Contribution

It combines domain adaptation and semi-supervised learning techniques for improved functional tissue unit segmentation at the cellular level.

Findings

Achieves state-of-the-art segmentation performance.

Effectively minimizes domain gap and class imbalance.

Demonstrates robustness across HPA and HubMAP datasets.

Abstract

We present a new method for functional tissue unit segmentation at the cellular level, which utilizes the latest deep learning semantic segmentation approaches together with domain adaptation and semi-supervised learning techniques. This approach allows for minimizing the domain gap, class imbalance, and captures settings influence between HPA and HubMAP datasets. The presented approach achieves comparable with state-of-the-art-result in functional tissue unit segmentation at the cellular level. The source code is available at https://github.com/VSydorskyy/hubmap_2022_htt_solution