Enforcing Morphological Information in Fully Convolutional Networks to Improve Cell Instance Segmentation in Fluorescence Microscopy Images

TL;DR

This paper introduces a novel cell segmentation method using a deep distance transformer within a U-Net framework, improving accuracy in challenging fluorescence microscopy images by incorporating morphological information.

Contribution

It proposes a deep distance transformer to enforce morphological learning in a U-Net, enhancing cell segmentation performance in microscopy images.

Findings

Performance boost over traditional U-Net models.

Effective incorporation of morphological information.

Potential for improved cell tracking and analysis.

Abstract

Cell instance segmentation in fluorescence microscopy images is becoming essential for cancer dynamics and prognosis. Data extracted from cancer dynamics allows to understand and accurately model different metabolic processes such as proliferation. This enables customized and more precise cancer treatments. However, accurate cell instance segmentation, necessary for further cell tracking and behavior analysis, is still challenging in scenarios with high cell concentration and overlapping edges. Within this framework, we propose a novel cell instance segmentation approach based on the well-known U-Net architecture. To enforce the learning of morphological information per pixel, a deep distance transformer (DDT) acts as a back-bone model. The DDT output is subsequently used to train a top-model. The following top-models are considered: a three-class (\emph{e.g.,} foreground, background and cell border) U-net, and a watershed transform. The obtained results suggest a performance boost over traditional U-Net architectures. This opens an interesting research line around the idea of injecting morphological information into a fully convolutional model.