Keywords: AI Diffusion Models, Diffusion Modeling
Motivation: Kolmogorov-Arnold Network (KANs) is an emerging deep learning technique and has a high potential to further improve the performance of deep learning based microstructural modeling for diffusion MRI.
Goal(s): Achieve accurate parameter estimation for DTI and NODDI models, enhancing efficiency and accuracy beyond that of conventional deep learning methods.
Approach: DiffKAN is proposed to map diffusion data to model parameters using convolutional KANs, with network parameters optimized by minimizing difference between predicted and reference parameter values.
Results: DiffKAN achieves lower mean absolute errors for both DTI and NODDI metrics compared to conventional CNN, UNet, and Vision Transformer models, demonstrating significant gains in accuracy.
Impact: DiffKAN’s efficient KAN-based architecture offers a pathway to accurate diffusion MRI modeling and analysis, significantly lowering computational burdens. DiffKAN might transform the clinical adoption of diffusion MRI, allowing for more widespread use in diagnostics by providing more accurate microstructural mapping.
How to access this content:
For one year after publication, abstracts and videos are only open to registrants of this annual meeting. Registrants should use their existing login information. Non-registrant access can be purchased via the ISMRM E-Library.
After one year, current ISMRM & ISMRT members get free access to both the abstracts and videos. Non-members and non-registrants must purchase access via the ISMRM E-Library.
After two years, the meeting proceedings (abstracts) are opened to the public and require no login information. Videos remain behind password for access by members, registrants and E-Library customers.
Keywords