Imaging Bootcamp

Imaging blood vessels is a part of research into all sorts of conditions, from stroke to Alzheimer’s. Analysing these data requires a process called blood vessel segmentation where blood vessels are identified from non-vessel tissue. With large volumes of imaging data, segmentation needs to be automated using computer algorithms. For this to be effective, algorithms first need to be trained on existing data. This currently relies on proprietary software where training data isn’t openly available or open-source software where training data isn't comprehensively labelled with what's what. Researchers now present MiniVess, a dataset of 70 3D image volumes of rodent brain blood vessels (a selection is pictured), imaged using two-photon fluorescence microscopy and labelled in detail. This was created using U-Net, a type of deep learning called a convolutional neural network. It is publicly available and will help train algorithms to analyse future blood vessel imaging data in brain research.

Written by Lux Fatimathas

Lux Fatimathas

Writer

Having pursued science through a career in research, communications and publishing, she is currently a writer and editor for a healthcare marketing agency and also juggles freelance science writing.

• Image from work by Charissa Poon and colleagues
• Sunnybrook Research Institute, Physical Sciences Platform, Toronto, Canada
• Image originally published with a Creative Commons Attribution 4.0 International (CC BY 4.0)
• Published in Scientific Data, March 2023