QSA is pleased to share that a paper by Dr Daniele Ravi and QSA colleagues on "DeepBrainPrint: A Novel Contrastive Framework for Brain MRI Re-Identification" has been accepted at the "Medical Imaging with Deep Learning" conference (MIDL 2023). Read the pre-print here.
DeepBrainPrint is a novel contrastive learning framework to learn compact representations of brain MRI scans, designed specifically to identify the same individual across multiple visits and scanners from medical imaging datasets.
Experimental results show superior performance over existing biometric verification methods, demonstrating strong re-identification robustness across multiple scanner manufacturers and field strengths.