Presentation
13 March 2024 Multispectral brain image registration and segmentation
Author Affiliations +
Abstract
This research aims to deal with intraoperative multispectral images taken from brain tumour surgeries to investigate the diagnostic and guidance potential of MSI. These images were registered by feature-based (SIFT, PFN), intensity-based (LK) and machine learning (RANSAC-Flow) methods and classified via a CNN and Transformer model using anatomical labels. Based on the results from some initial training, MSI could achieve 95% overall accuracy. After labelling and registration are completed, a brain surgery dataset can be built to support intraoperative decision making.
Conference Presentation
© (2024) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Zepeng Hu, Giulio Anichini, Stamatia Giannarou, Kevin O'Neill, and Daniel Elson "Multispectral brain image registration and segmentation", Proc. SPIE PC12831, Advanced Biomedical and Clinical Diagnostic and Surgical Guidance Systems XXII, PC1283104 (13 March 2024); https://doi.org/10.1117/12.3001385
Advertisement
Advertisement
KEYWORDS
Image registration

Brain

Image segmentation

Multispectral imaging

Neuroimaging

Microscopes

Education and training

Back to Top