ISMRM25 - Deep Learning for Image Enhancement: Part III

ISMRM & ISMRT Annual Meeting & Exhibition • 10-15 May 2025 • Honolulu, Hawai'i

Deep Learning for Image Enhancement: Part III

Digital Poster

AI & Machine Learning

Wednesday, 14 May 2025

Exhibition Hall

15:45 - 16:45

Session Number: D-40

No CME/CE Credit

		Computer Number: 33 3846. Consistently denoising 3D MR images using 2D neural networks View Presentation Video M. Haas, M. Herbst Bruker BioSpin GmbH & Co. KG, Ettlingen, Germany Impact: The presented method enables consistent high-dimensional denoising using 2-dimensional convolutional networks. Thus, processing of large MRI datasets becomes possible on standard workstations without the need of expensive computer hardware or connection to a remote server.
		Computer Number: 34 3847. AI-Accelerated Brain MRI: 30% Faster Scans with Uncompromised Diagnostic Accuracy for Aging Populations W. Gu, C. Yang, Q. zhang, S. Cai, D. Wu, Y. Dai Shanghai Punan Hospital of Pudong New Area, Shanghai, China Impact: ACS dramatically reduces brain MRI scan time while maintaining diagnostic accuracy. This offers a practical, time-efficient alternative for routine neuroimaging, particularly in elderly populations and resource-limited settings, enhancing patient comfort and clinical workflow without sacrificing diagnostic quality.
		Computer Number: 35 3848. DiffKAN: Convolutional Kolmogorov-Arnold Networks for Improved Diffusion MRI Microstructural Modeling View Presentation Video Y. Chen, Z. Li, Y. Wang, Z. Li, J. Zheng, H. Yang, M. Liu, Q. Tian Hangzhou Dianzi University, Hangzhou, China 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.
		Computer Number: 36 3849. Enhanced Diffusion MRI of Infant Brain at 0.35 Tesla Using a Self-Training Two-Stage Framework Z. Chen, Y. Ding, H. Xu, L. Zhao, H. Zhang, D. Wu Zhejiang University, Hangzhou, China Impact: The proposed method improved the image quality of DWI at low-field without collecting paired LF-HF data. It may promote accurate diagnosis using low-field MRI for DWI.
		Computer Number: 37 3850. Feasibility study of Attention-Guided Pix2Pix GAN for the synthesizing of enhanced cerebral vascular images in 7 Tesla MRI J. Tan, Z. Zhen, Q. Wang, W. Chen, Z. Wang, W. Chen Department of Medical Engineering, First Affiliated Hospital of Army Medical University, Chongqing, China Impact: The model we proposed can generate high-quality 7T magnetic resonance cerebrovascular enhancement images, which makes it possible to diagnose cerebral vessels without the use of contrast agents.
		Computer Number: 38 3851. Learning temporal characteristics in multi-contrast MR images with self-supervision: An application to accelerating quantitative T2 mapping View Presentation Video L. Umapathy, H. Pei, N. Ben-Eliezer, D. Sodickson, L. Feng NYU Grossman School of Medicine, New York, United States Impact: An understanding of underlying temporal characteristics of tissues with vision transformers can help with intelligent design of current multi-contrast data acquisition schemes.
		Computer Number: 39 3852. Self-supervised deep learning model for denioising and distortion correction in accelerated echo planar imaging View Presentation Video J. Kim, B. Kang, H. Park, H. Seo Korea Institute of Science and Technology (KIST), Seoul, Korea, Republic of Impact: The proposed self-supervised approach achieves a significant advancement by enabling simultaneous denoising and distortion correction in accelerated EPI without ground truth images, thereby enhancing image quality in accelerated imaging.
		Computer Number: 40 3853. MRI-Enhanced Generative Model for Alzheimer's Disease: Converting 18F-FDG PET to 18F-AV-45 PET View Presentation Video X. Fu, Y. Jin, H. Shao, H. Liu, Y. Zhang, N. Zhang, H. Zheng, D. Liang, J. Liu, Z. Hu Research Center for Medical AI, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China Impact: This method significantly enhances diagnostic efficiency in Alzheimer's disease by enabling quick, cost-effective image generation, reducing reliance on expensive, short-lived tracers, and providing accessible support for clinicians, ultimately advancing multi-modal medical imaging practices in neurodegenerative diseases.
		Computer Number: 41 3854. 3D fetal head pose estimation from MRI navigators with equivariant networks View Presentation Video R. Muthukrishnan, B. Billot, B. Gagoski, M. Firenze, M. Soldatelli, P. E. Grant, P. Golland Massachusetts Institute of Technology, Cambridge, United States Impact: Our method demonstrates the promise of fetal head pose estimation and opensnew possibilities for dynamic prescription of the imaging plane that tracks thefetal head in real time.
		Computer Number: 42 3855. Advanced ZTE MR Lung Imaging: A Deep Learning Approach to Enhance SNR and Reduce Artifacts View Presentation Video J. de Arcos, S. Mandava, M. Lebel, F. Wiesinger, C. Cretu, P. Wielopolski, J. Hernández Tamames, P. Ciet GE HealthCare, Chalfont Saint Giles, United Kingdom Impact: The DL-ZTE model significantly enhances lung MRI quality by reducing artifacts and improving SNR while maintaining anatomical accuracy, facilitating more accurate and reliable clinical assessments of parenchymal abnormalities in ZTE lung imaging applications.
		Computer Number: 43 3856. Spatiotemporal 4D-UNet for Physics Consistent Super-Resolution and Denoising of 4DFlow-MRI View Presentation Video A. Ghazipour, A. Kazemi, A. Amini University of Louisville, Louisville, United States Impact: Our model can improve 4D-Flow MRI data that has been hindered by noise, artifacts, and lower resolution. As a result, hemodynamic parameters that are critical for diagnosing various disease can be more accurately measured
		Computer Number: 44 3857. A Dual-Stage Denoising Method Based on Zero-Shot Learning for Low-Field MRI View Presentation Video Y. Li, S. Liu, Y. Liu, M. Lyu Shenzhen Technology University, Shenzhen, China Impact: The framework of our proposed dual-stage denoising method is plug-and-play for various existing denoising models and generally enhances their performance.
		Computer Number: 45 3858. Specialized Coil Informed Deep Learning for High SNR Carotid Imaging View Presentation Video L. Zeng, M. Lu, Y-C Hsu, M. Keushkerian, K-L Nguyen, K. Johnson, M. Altbach, H. D. Morris, J. K. DeMarco, V. Deshpande, D. Mitsouras, D. Saloner, S. McNally, S-E Kim, J. Roberts, R. Hadley, G. Treiman, D. Parker, D. Li, Y. Xie Cedars-Sinai Medical Center, Los Angeles , United States Impact: With the use of our DL model, high SNR images are achievable with standard head-neck coils, which may help radiologists to be more confident and efficient in evaluating carotid plaque characteristics.
		Computer Number: 46 3859. Simultaneous reduction of noise and motion artifacts in brain MRI using deep learning. View Presentation Video I. Muro, S. Shibukawa, K. Usui AIC YAESU CLINIC, Tokyo, Japan Impact: By 36,000 pairs of training data, we were able to increase the accuracy of the learning process. The advantage of this method is that it is post-processing and can be used regardless of the equipment or imaging method.
		Computer Number: 47 3860. Enhanced Sodium Imaging at 3T MRI Using BSRGAN: Achieving High SNR and Spatial Resolution View Presentation Video S. Kim, S. Oh, S-Y Kim, J. Hwang, C. Y. Lim, Y. Sim, E. Kim, E. S. Ko, W. K. Jeong, S. T. Kim, J. H. Lee Department of Radiological Science, Daewon University College, Jecheon, Korea, Republic of Impact: BSRGAN-based reconstruction of sodium images achieved a 1.4-fold increase in signal-to-noise ratio and a twofold improvement in spatial resolution at 3T, significantly enhancing image quality and reducing acquisition time.

Similar Session(s)