| 08:15 | | Introduction |
| 08:27 |  | 0383. MRI Reconstruction with Learnable High Order Fast Fourier Transform (HOFFT) Kernels D. Abraham, M. Nishimura, Z. Shah, J. Pauly, K. Setsompop
Stanford University, Stanford, United States Impact: Our method will help reduce the computational burden of high order phase correction in MRI. This opens the door to rapid 3D encoding schemes such as cones, MRF, and time-resolved imaging, potentially turning previously intractable problems to feasible ones. |
| 08:39 |  | 0384. Smooth operators: exploring B-splines as learnable non-linear activation functions for complex-valued MRI reconstruction M. Terpstra, C. A. van den Berg
UMC Utrecht, Utrecht, Netherlands Impact: Spline-based complex-valued neural networks might improve image quality and enable further acceleration of MRI acquisitions. These results can better help to diagnose patients based on MRI exams while improving patient comfort by reducing the MRI acquisition time.
|
| 08:51 |  | 0385. Robust Multi-Contrast MR Reconstruction Based on Disentangled Representation Learning-Embedded Deep Unrolling Z. Xue, C. Hu
National Engineering Research Center of Advanced Magnetic Resonance Technologies for Diagnosis and Therapy, School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai, China Impact: This DRL-assisted reconstruction approach has the potential to serve as
a universal model for multi-contrast MR data. |
| 09:03 |  | 0386. Noise Analysis in Physics-based Deep Unrolled Network Reconstructions O. Dalmaz, A. Desai, A. Chaudhari, B. Hargreaves
Stanford University, Stanford, United States Impact: A practical tool facilitates noise analysis in physics-based deep learning reconstructions, so that image SNR can be assessed more objectively between reconstructions and patient scans. This would encourage researchers to develop more robust and trustworthy algorithms. |
| 09:15 |  | 0387. Self-Consistency-Driven Test-Time Prompt Tuning for All-in-One MR Reconstruction Model Z-X Cui, T. Xie, X. Wang, W. He, C. Liu, Q. Zhu, Y. Liu, J. Cheng, Y. Zhou, D. Liang
Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China Impact: This model can adapt to all MRI scenarios, facilitating simplified installation and maintenance. |
| 09:27 |  | 0388. Federated Visual Autoregressive Transformers for Collaborative Model Training in MRI Reconstruction V. Nezhad, G. Elmas, T. Cukur
Bilkent University, Ankara, Turkey Impact: High-fidelity image generation achieved by FedVAT enables imaging sites to collaboratively train MRI reconstruction models with divergent architectures. Avoidance of architectural constraints combined with reliable generalization can facilitate applications that suffer from data scarcity, such as assessment of rare diseases. |
| 09:39 |  | 0389. Bilevel Optimized Implicit Neural Representation for Scan-Specific Accelerated MRI Reconstruction H. Yu, J. Fessler, Y. Jiang
University of Michigan, Ann Arbor, United States Impact: By automatically optimizing hyperparameters for scan-specific deep learning, our method reconstructs accelerated MRI scans across diverse protocols with superior image quality. It avoids reliance on training data and complicated task-dependent tuning, enhancing the clinical applicability of deep learning in MRI. |
| 09:51 |  | 0390. Hybrid Learning: A Novel Combination of Self-Supervised and Supervised Learning for MRI Reconstruction without High Quality Training Reference H. Pei, Y. Wang, H. Chandarana, L. Feng
New York University Grossman School of MedicineNew York University Grossman School of Medicine, New York, United States Impact: This study proposes a novel hybrid learning strategy to address challenges when obtaining high-quality reference data is difficult, which enables more accurate reconstruction at higher acceleration rates, which is beneficial in various applications where only low-quality reference images are available . |
| 10:03 |  | 0391. Advancing Self-Supervised Learning for Highly Accelerated MRI Reconstruction Through Parallel Imaging Consistency Y. U. Alcalar, C. Zhang, M. Akçakaya
University of Minnesota, Minneapolis, United States Impact: This work proposes an improved training strategy for self-supervised MRI
reconstruction by applying well-designed perturbations to input images. This
ensures alignment with parallel imaging techniques and reduces aliasing
artifacts, achieving visible improvements at high acceleration rates. |