Image Segmentation with Topological Priors

TL;DR

This paper demonstrates that integrating topological priors into deep neural network training, specifically with a UNet model, improves segmentation accuracy and topological correctness, especially in fine-scale structures.

Contribution

It introduces a method to incorporate topological priors during training, enhancing segmentation performance and topological accuracy over traditional approaches.

Findings

Topological priors improve segmentation accuracy.

Incorporating topological info reduces Betti number errors.

Topological training outperforms classical UNet on the ISBI EM dataset.

Abstract

Solving segmentation tasks with topological priors proved to make fewer errors in fine-scale structures. In this work, we use topological priors both before and during the deep neural network training procedure. We compared the results of the two approaches with simple segmentation on various accuracy metrics and the Betti number error, which is directly related to topological correctness, and discovered that incorporating topological information into the classical UNet model performed significantly better. We conducted experiments on the ISBI EM segmentation dataset.