Cape Town - 2026 ISMRM-ISMRT Annual Meeting and Exhibition

Registration, Atlases, and Motion

Digital Poster

Analysis Methods

Monday, 11 May 2026

Digital Posters Row I

08:20 - 09:15

Session Number: 368-01

No CME/CE Credit

Similar Sessions

This session will highlight recent advances in image registration, with an emphasis on AI-based methods. Contributions will span atlas construction and motion-related applications, reflecting current trends and open challenges in the field.

Skill Level: Intermediate

		368-01-001. SGDR-Net: A Self-Guide Deformable Registration Network in DCE-MRI breast imaging Yichao Zhou, Shihui Zhou, Bing Han, LIJUN ZHANG Canon Medical Systems (China), Beijing, China Impact: A highly accurate deep learning registration method has been proposed. The proposed method can reduce the impact of motion deformation on subtraction images, improve diagnostic accuracy, and reduce patient repeat scans caused by motion.
		368-01-002. Direct deformation estimation with recurrent inference machines for longitudinal MRI Jonas Maes, Jana Osstyn, Matthan Caan, Arnold den Dekker, Jan Sijbers University of Antwerp, Antwerp, Belgium Impact: Longitudinal MRI becomes increasingly important³, expressing the need for a more time-efficient and user-friendly follow-up imaging approach. Our RIM-augmented Delta-MRI enables substantially shorter follow-up scans, thereby contributing to the reduction of waiting lists and a more cost-effective patient follow-up tool.
		368-01-003. DeepSUM-DWI: A combined registration and super-resolution model for diffusion-weighted imaging Merlin Fair, Carolina Daniells Zaldívar, Paola Ocampo Luna, Luis Concha, María Guadalupe García-Gomar Universidad Nacional Autonoma de Mexico, Queretaro, Mexico Impact: Typical diffusion weighted imaging acquisitions are often limited in spatial resolution and image SNR. The proposed model provides the potential for higher apparent spatial resolution, even from a low-resolution input, while also aiding SNR and potential subject motion between repetitions.
		368-01-004. Population-Level 4D Cardiac Atlas for Deep Phenotyping and Cardiovascular Disease Diagnosis Guangming Wang, Qirong Li, Lizhen Lan, Yajing Zhang, Mo Yang, Qing Li, Tianxing He, Yan Li, Chengyan Wang Human Phenome Institute, Shanghai, China Impact: This study establishes a large-scale 4D cardiac atlas from 50,090 UK Biobank (UKB) CMR datasets, enabling comprehensive quantification of cardiac shape and motion. It advances population-level cardiac phenotyping and enhances disease diagnosis by capturing spatiotemporal representations beyond conventional 3D models.
		368-01-005. Motion Compensation in Dynamic Free Breathing Pulmonary UTE MRI Reconstruction for Neonatal Subjects with Bulk Motion Sabbir Ahmed, Andrew Hahn, Abhilash Kizhakke Puliyakote, Sean Fain, Mathews Jacob University of Iowa, Iowa City, United States of America Impact: Dynamic motion compensated reconstruction of pulmonary MRI allows visualizing the true dynamics and functional imaging of the lungs.
		368-01-006. Improving Multicontrast Arterial Centerline Alignment Between TOF and SNAP MRI Using Landmark-Based Affine Transformations Abrar Uddin University of Utah, Salt Lake City, United States of America Impact: Our automated TOF-to-SNAP MRA conversion tool replicates manual modeling accuracy in minutes, eliminating hours of labor. It enables visualization of arteries often omitted in SNAP images, allowing faster, more complete vascular centerlines and improving efficiency and accessibility in cerebrovascular research.
		368-01-007. Integrated Optimization of Distortion Correction and Functional-Anatomical Image Co-registration in fMRI Joseph Hutter, Qingfei Luo, Xiaohong Joe Zhou University of Illinois Chicago, Chicago, United States of America Impact: Combining distortion correction and functional-anatomical image co-registration into one optimization step improves the accuracy of distortion correction and image registration in fMRI.
		368-01-008. Towards Accelerated Motion-Resolved 4D Cartesian Imaging Using Compressed Sensing with Deep Learning Motion Estimation Majd Helo, Aya Ghoul, Daniel Amsel, Cornelius Eichner, Marcel Dominik Nickel, Thomas Küstner University Hospital of Tuebingen, Tuebingen, Germany Impact: The self-navigated, motion-resolved Cartesian 4D-MRI integrates 3D deep-learning registration with CS reconstruction to improve consistency and vessel conspicuity in abdominal imaging and facilitates radiotherapy planning, enhancing patient comfort, especially in RT-settings where flat treatment tables can cause pain and discomfort.
		368-01-009. Label‑Free Principal Strain from Cine SSFP via Unsupervised Deformable Registration and Myocardial‑Aware Training Ali Badran, Muhannad Abdallah, Osama Mahmoud, Ahmed Hassan, Mishkat Habib, Tamer Basha, Ahmed Gharib, Ahmed Abdelfadeel, Khaled Abd-Elmoniem Systems and Biomedical Engineering Department, Faculty of Engineering, Cairo University, Giza, Egypt Impact: Unsupervised, label‑free cine CMR registration yields dense, pixel‑wise principal strain, addressing feature-tracking (FT) boundary bias and supervised methods’ label dependency, and opening a scalable path to standardized myocardial function maps across sites and vendors.
		368-01-010. Concurrent Slice-Level Motion Monitoring and Volume-Level Prospective Motion Correction in Functional MRI Joshua Auger, Musa Tunc Arslan, Hongli Fan, Simon Warfield Boston Children's Hospital and Harvard Medical School, Boston, United States of America Impact: This work combines volume-level PMC to maintain FoV alignment with concurrent slice-level monitoring to characterize residual motion, allowing for real-time, motion-aware acquisition decisions to ensure sufficient fMRI quality in motion-prone populations, such as pediatric epilepsy.
		368-01-011. Joint Optimization of Acquisition, Reconstruction, and Registration for Maximizing Motion Estimation in Dynamic Cardiac MRI George Yiasemis, Jan-Jakob Sonke, Jonas Teuwen the Netherlands Cancer Institute, Amsterdam, Netherlands Impact: This end-to-end framework demonstrates that learning adaptive sampling, reconstruction, and registration jointly leads to superior motion estimation from undersampled dynamic MRI. It establishes an initial foundation for task-driven, real-time dynamic imaging applicable MRI-guided workflows.
		368-01-012. A microstructure-informed common coordinate framework for the macaque brain Ricardo Gonzales, Ting Gong, Jingjing Wu, Jasmine Shao, Elissa Bell, Suzanne Haber, Yaël Balbastre, Anastasia Yendiki Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital and Harvard Medical School, Charlestown, United States of America Impact: A microstructure-informed macaque brain template unifies structural, orientational, and cellular contrasts, improving cross-modal registration. Within the BRAIN Initiative CONNECTS program, it forms a foundation for linking MRI and microscopy, enabling multiscale investigations of non-human primate brain organization and connectivity.
		368-01-013. Enhanced Cardiac DCE MRI Quantification with Multi-Level Motion Correction Xinqi Li, Li-Ting Huang, Xinheng Zhang, Yi Zhang, Thoralf Niendorf, Min-Chi Ku, Qian Tao, Hsin-Jung Yang Cedars-Sinai Medical Center, Los Angeles, United States of America Impact: Incorporating multi-level motion correction significantly improves the reliability of cardiac DCE quantification, by minimizing motion-induced bias in voxel-wise pharmacokinetic fitting. This approach enhances the accuracy and reproducibility of myocardial perfusion assessment, supporting robust physiological interpretation and longitudinal monitoring in CMR.
		368-01-014. Fully Automatic Slice-to-Volume MRI–Histology Registration: A Pilot Study on Myelin Staining Sutatip Pittayapong, Simon Hametner, Beata Bachrata, Verena Endmayr, Christian Menard, Wolfgang Bogner, Romana Höftberger, Guenther Grabner Carinthia University of Applied Sciences, Klagenfurt, Austria Impact: This work enables automatic registration of single histology slices to 3D MRI volumes without manual input or auxiliary data. It links microscopic and macroscopic information, improving MRI biomarker validation and advancing neuroimaging research for applications in microstructural and disease analysis.
		368-01-015. Hybrid AI–Physics Framework for 3D Heart Motion Reconstruction Using Differential Scattering Parameters Ettore Flavio Meliado, Vladislav Koloskov, Cornelis van den Berg, Bart R Steensma University Medical Center Utrecht, Utrecht, Netherlands Impact: Tracking cardiac biomarkers such as stroke volume and ejection fraction in home or outpatient settings is crucial for cardiovascular disease management. This work develops and validates a radio-frequency sensing (RFS) method using a wearable antenna array, benchmarked against MRI-derived biomarkers.
		368-01-016. Low-field, high-gradient diffusion NMR on in vivo mouse brain Nathan Williamson, Rea Ravin, Natalia Gudino, Emily Long, Teddy Cai, Peter Basser The Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., Bethesda, United States of America Impact: We demonstrate the use of a portable low-field, single-sided magnet with a strong static gradient to measure spin echo signals at unprecedented diffusion weightings in the in vivo rodent brain. Attenuation curves show evidence of motional averaging or localization.