Skip to content
← Projects

ML / AI · CISS 2025, Published

Mouse Brain Cell Segmentation in Fluorescence Microscopy

A deep learning segmentation pipeline for high-noise fluorescence microscopy images, comprising a CNN architecture and a custom preprocessing routine for automated cell boundary detection.

Fluorescence

Modality

Cell boundary

Task

CISS 2025

Venue

Problem

Fluorescence microscopy of mouse brain tissue is dominated by Poisson noise and uneven illumination, conditions under which off-the-shelf segmentation models perform poorly.

Approach

I implemented CNN-based segmentation paired with custom preprocessing, including denoising and normalization tailored to the imaging conditions, alongside a reproducible training loop and qualitative validation against expert annotations.

Results

The pipeline enabled automated cell boundary detection and was accepted at CISS 2025. It has since become the laboratory's reference implementation for downstream microscopy work.

Stack

OpenCVPyTorchCNNs