This repository contains the official implementation for the paper:
Vision-Language-Guided Granularity Alignment and Balanced Fusion for Multimodal Brain Data
Before running GABI, please ensure you have the necessary data and a properly configured environment.
GABI is trained and evaluated on publicly available medical datasets.
You will need to obtain access to the following datasets:
-
ADNI (Alzheimer's Disease Neuroimaging Initiative)
https://adni.loni.usc.edu/ -
OASIS-1 (Open Access Series of Imaging Studies)
https://sites.wustl.edu/oasisbrains/ -
ABIDE I (Autism Brain Imaging Data Exchange)
https://fcon_1000.projects.nitrc.org/indi/abide/abide_I.html -
AIBL (Australian Imaging, Biomarkers and Lifestyle Study of Ageing)
https://aibl.csiro.au/