Work

Automated Generation of Neuron Segmentation Masks From Fluorescence Microscopy as a Foundation for Phase‐Contrast Deep Learning Applications

Year: 2026

Type: article

Abstract: ABSTRACT Accurate segmentation of neurons in microscopy images of neuronal cultures is crucial for research on neuro‐degenerative diseases and neurotoxicity. Manual annotation is time‐consuming, subjective and inconsistent across experts, motivating deep learning (DL) approaches that require high‐quality annotated training datasets. Neuronal cultures can be imaged using either phase‐contrast or fluorescence microscopy. While live neurons are easily detected in fluorescence images, this modality requires staining that affects cell viability, whereas phase‐contrast imaging is non‐invasive but makes neuron identification more challenging. We present a robust computer vision pipeline for automatically generating expert‐level neuron segmentation masks from fluorescence microscopy images. These masks are intended to facilitate the efficient training of robust DL models for neuron identification in phase‐contrast images. Our method leverages paired fluorescence and phase‐contrast images captured simultaneously at the same magnification and coordinates, ensuring alignment across modalities, so that once trained, the models can segment neurons in phase‐contrast images alone, eliminating the need for fluorescence staining. We validate our approach by comparing automatically generated masks with expert annotations using multiple agreement metrics, including intersection over union (IoU), Matthew's correlation coefficient (MCC) and the Jaccard index, covering both identification and instance segmentation. Our results demonstrate that a classical computer vision algorithm can generate neuron segmentation masks with reliable instance separation, achieving expert‐level performance and enabling downstream applications such as non‐invasive cell viability analyses via convolutional neural networks, thus contributing to the field of neuro‐biological research.

Source: IET Image Processing

DOI: 10.1049/ipr2.70511

Authors: Gerard Villarroya‐Piqué, Víctor M. González, Esther Serrano‐Pertierra, M. Teresa Fernandez‐Sanchez, Antonello Novelli, Angel Rio‐Alvarez

Institution: Universidad de Oviedo

Language: English

Topic: Cell Image Analysis Techniques

Subfield: Biophysics

Field: Biochemistry, Genetics and Molecular Biology

Domain: Life Sciences

SDG: Good health and well-being

Open Access status: gold

Estimated APC paid: $2,740