Cape Town - 2026 ISMRM-ISMRT Annual Meeting and Exhibition - Super-Resolving MRI: Methods and Applications

Super-Resolving MRI: Methods and Applications

Digital Poster

Analysis Methods

Wednesday, 13 May 2026

Digital Posters Row J

13:40 - 14:35

Session Number: 569-03

No CME/CE Credit

Similar Sessions

The session showcases image enhancement techniques in MRI with a particular focus on super-resolution tasks.

Skill Level: Basic,Intermediate,Advanced

		569-03-001. Deep Learning Reconstruction of Isotropic Structural MRI from Routine 2D Clinical Scans JiaChen Sun, Yayan Yin, Boyan Xu, Jie Lu Xuanwu Hospital, Capital Medical University, Beijing, China Impact: This method reconstructs isotropic structural images directly from routine 2D acquisitions, preserving anatomical and fidelity while eliminating the need for separate high-resolution 3D scans. It improves workflow efficiency and accessibility for quantitative and functional MRI studies in clinical settings.
		569-03-002. 3D Generative AI Super-Resolution for Free-Running Cardiac MRI at 0.6 T Omer Demirel, Dinghui Wang, Spencer Waddle, Vincenzo Scialò, Enas Ahmed, Tzu Cheng Chao, Jouke Smink, Jérôme Yerly, Robert Holtackers, Jacinta Browne, Tim Leiner Philips North America Clinical Science, Rochester, United States of America Impact: This study demonstrates that deep learning-based 3D super-resolution enhances free-running 5D cardiac MRI at 0.6T acquired in half the scan time. The approach improves image sharpness and spatial details at low field strength, supporting faster free-running whole-heart imaging.
		569-03-003. Deep Learning-based Super-Resolution for BLADE MRI using Simulated and Native Training Datasets Keerthi Prabhu M, Punith Bidarakka Venkategowda, Asha KumaraSwamy Kuppe, Kun Zhou, Thomas Benkert, Seung Su Yoon, Marcel Dominik Nickel Magnetic Resonance, Siemens Healthineers, Bangalore, India Impact: The results demonstrated that while BLADE-simulated-TSE datasets can act as a useful pretraining strategy, BLADE datasets remain essential to further improve reconstruction performance. This highlights the importance of dataset-specific training, providing a roadmap for extending DL-based SR to non-Cartesian acquisitions.
		569-03-004. Enhancing Low-Field MRI Image Quality for Nipah Virus Infection Imaging using Deep Learning Ajay Sharma, Ivan Etoku Oiye, Russell Byrum, Michael Holbrook, Yu Cong, Claudia Calcagno, Venkatesh Mani, Sairam Geethanath Johns Hopkins University School of Medicine, Baltimore, United States of America Impact: This study aims to improve the use of LF-MRI for Nipah virus induced brain lesion detection. The optimal T2w parameters have been identified, and the super-resolution method improves image quality for possibly monitoring infectious disease (NiV) outbreaks.
		569-03-005. Attention-Enhanced Latent Diffusion for MRI Super-Resolution Peijiang Ma, Zhongsen Li, Juanhua Zhang, Kaihan Yang, Haoding Meng, Rui Li Tsinghua University, Beijing, China Impact: By adopting the latent diffusion model, our approach not only facilitates high-quality MRI super-resolution but also substantially reduces training resource requirements, which is beneficial for promoting the wider use of MRI super-resolution techniques.
		569-03-006. Validation and Feasibility of Fast Knee MRI Using a Deep Learning-Assisted 3D Iterative Image Enhancement System xiaoqing Wu, Xi Zhu, Jie Shi, Jing Ye, Wennuo Huang, Wei Xia Dalian Medical University, Dalian, China Impact: The results support the feasibility of DL-3DIIE as a practical solution for achieving rapid, high-quality knee MRI without hardware modification.
		569-03-007. Fast undersampled dynamic MRI reconstruction using explicit representation learning with Gaussian splatting Maarten Terpstra, Cornelis van den Berg University Medical Center Utrecht, Utrecht, Netherlands Impact: Quickly obtaining high-quality MRI from accelerated acquisitions is important to mitigate motion artifacts, maintain patient comfort, and improve clinical efficiency. We show that explicit representation learning using Gaussian splatting enables high-quality cardiac MRI, while directly encoding the underlying tissue properties.
		569-03-008. Revolutionizing MRI Reformatting: ESRGAN Transfer Learning for Through-Plane Resolution Enhancement YASHWANT KURMI, Malvika Viswanathan, Zhongliang Zu Vanderbilt University Institute of Imaging Science, Vanderbilt University Medical Center, Nashville, United States of America Impact: The proposed Through-Plane Resolution Enhancement (T-PRE) framework enhances through-plane resolution in multi-slice MRI reformatting by combining slice profile inversion with deep learning-based restoration (MRI-ESRGAN). It achieves superior image quality, improves structural fidelity, shortens acquisition time.
		569-03-009. Beyond the Plane: Multi-Planar Super-Resolution MRI Improves 3D Spinal Cord Lesion Assessment Tim Emmenegger, Steffen Franz, Michael Adam, Anna Duguid, Patrick Will, Simon Schading-Sassenhausen, Bernhard Strasser, Sabina Frese, Wolfgang Marik, Wolfgang Bogner, Simon Robinson Medical University of Vienna, Vienna, Austria Impact: Spinal cord injury can cause focal lesions that disrupt ascending and descending pathways. To fully realize the potential of emerging 3D lesion assessments, isotropic high-resolution imaging strategies robust to metal artefacts are essential to improve accuracy, sensitivity, and reproducibility.
		569-03-010. CONSIS-Net — Consistency-based Iterative Super-Resolution unrolled Network for low-field knee MRI Sebástian Ibarra, Steren Chabert, Ronal Coronado, Claudia Prieto Millennium Institute for Intelligent Healthcare Engineering - iHEALTH, Santiago, Chile Impact: This work demonstrates that a neural network integrating transformer-based super-resolution module with data-consistency steps can produce high-quality images from lower-resolution (4x times faster) acquisitions. The proposed framework improves image sharpness and structural fidelity without increasing scan duration or patient burden
		569-03-011. Spectrally-Optimized Implicit Neural Representations for Super-Resolving Cerebrovascular 4D Flow MRI Oliver Welin Odeback, Nerea González-Aranceta, Sebastiàn Jofre, Javier Bisbal, Edward Ferdian, Alistair Young, Pia Callmer, Mia Bonini, David Nordsletten, Susanne Schnell, Tobias Granberg, Alexander Fyrdahl, David Marlevi Karolinska Institutet, Solna, Sweden Impact: Using a spectrally optimized bias, INR enables accurate, rapid, full-field super-resolution recovery from standard-resolution 4D Flow MRI with settings transferable across patient anatomies.
		569-03-012. Non-Cartesian Super-resolution with Homodyne Partial Fourier Reconstruction Haneefah Brnawi, Michael Mendoza, Neal Bangerter, Peter Lally Imperial College London, London, United Kingdom Impact: We propose a practical solution for capturing high-resolution information of rapidly decaying signals, using a super-resolution approach and partial Fourier reconstruction. This technique extends k-space achieving a threefold resolution increase in one dimension with just two acquisitions.