ISMRM25 - Deep Learning for Image Enhancement: Part II

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

Deep Learning for Image Enhancement: Part II

Digital Poster

AI & Machine Learning

Wednesday, 14 May 2025

Exhibition Hall

15:45 - 16:45

Session Number: D-34

No CME/CE Credit

		Computer Number: 17 3830. QG-MoCo: Quality-Guided Coarse- and Fine-Grained Path Selection For MRI Motion Correction View Presentation Video F. Li, Z. Zhou, Y. Fang, J. Cai, Q. Wang ShanghaiTech Univerisity, Shanghai, China Impact: The quality-guided MoCo method effectively reduces 3D motion artifacts following the automatic selective path in different granularity for progressive correction. Emphasizes the benefits on model accuracy and efficiency by taking MoCo operations based on the routing strategy with quality consideration.
		Computer Number: 18 3831. Self-supervised multi-instance contrastive learning for reduction of cardiac bSSFP off-resonance artifacts View Presentation Video Z. Chen, Y. Emu, J. Gao, H. Chen, X. Tang, C. Hu Shanghai Jiao Tong University, Shanghai, China Impact: Obtaining corrupted-clean bSSFP image pairs is challenging, particularly with cardiac devices or high-field MR. Our proposed self-supervised approach mitigates off-resonance artifacts and provides a practical solution for reliable bSSFP cine imaging.
		Computer Number: 19 3832. Appearance Transformation Consistency Network for Registering 1H MRI and 129Xe MRI of the lungs View Presentation Video Z. Bao, S. Xiao, Z. Chen, X. Zhou State Key Laboratory of Magnetic Resonance and Atomic and Molecular Physics, National Center for Magnetic Resonance in Wuhan, Wuhan Institute of Physics and Mathematics, Innovation Academy for Precision Measurement Science and Technology, Chinese Academy of Sciences–Wuhan National Laboratory for Optoelectronics, Huazhong University of Science and Technology, 430071, wuhan, China Impact: Our results will impact radiologists and researchers by enabling precise lung image registration, facilitating new investigations into lung pathologies.
		Computer Number: 20 3833. Unsupervised Denoising Method for Multi-sequence Low-field MRI in Veterinary Imaging View Presentation Video J. Tang, W. Zu, J. Liu, C. Jin, Z. Zhang School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai, China Impact: In the absence of reference images, our method preserves the structural features of images while effectively reducing noise. It achieve noise reduction for low-field images in a shorter time and has great potential for improving low-field image quality improvement.
		Computer Number: 21 3834. FuseMorph: accurate and time-efficient MRI 3D T1 Image deformable registration with iterative search and deep learning View Presentation Video P-M Sun, T-Y Huang, T-C Chuang, Y-R Lin, H-W Chung Department of Electrical Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan Impact: This method improves MRI alignment accuracy and accelerates processing, offering a reliable tool for both research and clinical applications. It also enhances downstream tasks, such as VBM analysis, allowing them to be performed with greater speed and precision.
		Computer Number: 22 3835. MRCP-Specific Adaptive Volume Clipping: A Deep Learning Method for Automated Removal of Unnecessary Areas in MRCP images View Presentation Video Y. Sugimoto, N. Fujita, S. Funayama, S. Ichikawa, S. Goshima, Y. Terada University of Tsukuba, Tsukuba, Japan Impact: The MRCP-Specific Adaptive Volume Clipping method enhances MRCP examination efficiency by automatically removing unnecessary regions in maximum-intensity projection images of MRCP. This approach reduces manual workload, offering flexibility without precise segmentation, and improves workflow and diagnostic accuracy.
		Computer Number: 23 3836. Optimizing feature-based loss functions for AI Super-resolution of 7T Brain Diffusion MRI View Presentation Video D. Lohr, R. Werner University Hospital Hamburg-Eppendorf, Hamburg, Germany Impact: Our results show how feature-based loss functions need to be adapted to work well for SR models targeting MR diffusion data. We further demonstrate how such SR models may be trained using publicly available data, enabling reproducibility and application.
		Computer Number: 24 3837. Accelerated EPR Imaging Using Deep Learning Denoising View Presentation Video I. Canavesi, N. Viswakarma, B. Epel, A. McMillan, M. Kotecha O2M Technologies, LLC, Chicago, United States Impact: EPR images with physically enhanced deep learning techniques improve image SNR and reduce artifacts. This advancement can be translated to reduce acquisition time, reduce deposited power, and enable large object oxygen imaging, bringing EPRI one step closer to clinical translation.
		Computer Number: 25 3838. Deep Learning Image Denoising for In-Vivo Low-Field MRI Using Test-Time Training View Presentation Video D. Schote, C. Kolbitsch, L. Winter, A. Kofler Physikalisch-Technische Bundesastalt (PTB) Braunschweig und Berlin, Berlin, Germany Impact: Noise-free reference data for supervised training of low-field MRI denoising models does not exist. Using supervised pretraining on simulated data combined with self-supervised test-time training narrows the performance gap in low-field MRI denoising models when training and testing data differs.
		Computer Number: 26 3839. Multimodal Feature-Guided Diffusion Model for Low-Dose PET Image Denoising with MRI View Presentation Video Y. Jin, G. Lin, H. Liu, C. Zhou, X. Zhang, W. Fan, N. Zhang, H. Zheng, D. Liang, P. Cao, Z. Hu Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China Impact: The research can enhance the quality of low-dose PET imaging and reduce the radiation risk for patients, and it also inspires more feasible imaging solutions for the global health field, holding significant scientific and clinical importance.
		Computer Number: 27 3840. Denoising 7T Structural MRI with Conditional Generative Diffusion Models View Presentation Video B. Li, Y. Wang, Y. Liang, M. Carlson, P. DiGiacomo, H. Moein Taghavi, J. Maclaren, M. Bell, E. Mormino, V. Henderson, G. Zaharchuk, B. Rutt, W. Shao, M. Georgiadis, M. Zeineh Stanford University, Palo Alto, United States Impact: Our newly introduced 7T Conditional Diffusion Model (7TCDM) enables faster MRI acquisition by providing high-quality denoised images from shorter scans, increasing the feasibility of scanning patients in shorter times while preserving essential anatomical details.
		Computer Number: 28 3841. AI-Enhanced Super-Resolution for Metabolite MRI Imaging View Presentation Video E. Bjørkeli, J. T. Geitung, M. Esmaeili Akershus University Hospital, Lørenskog, Norway Impact: This improved SR approach significantly enhances metabolite map quality, offering clinicians a valuable tool for detailed neurological assessment.
		Computer Number: 29 3842. Applications of generative adversarial networks for super-resolution of cerebrovascular 4D Flow MRI View Presentation Video O. Welin Odeback, E. Ferdian, A. Young, J. Schollenberger, C. A. Figueroa, T. Granberg, A. Fyrdahl, D. Marlevi Karolinska Institute, Stockholm, Sweden Impact: This study highlights the potential of generative adversarial networks to enhance super-resolution in 4D Flow MRI, enabling more accurate intracranial flow assessments near vessel walls.
		Computer Number: 30 3843. Partial volume estimation from MRF acquisition using a deep learning approach View Presentation Video T. Ding, Y. Gao, Z. Xiong, F. Liu, M. Cloos, H. Sun The University of Queensland, Brisbane, Australia Impact: This study's self-supervised deep learning approach for partial volume estimation directly from MRF signals could improve diagnostic accuracy and streamline quantitative MRI processes. It opens avenues for real-time, artifact-resilient tissue characterization, potentially transforming clinical workflows and supporting patient-specific imaging studies.
		Computer Number: 31 3844. A Physics-Informed Deep Learning Method for Correcting Motion Artifacts in Brain MR Imaging View Presentation Video M. Safari, Z. Eidex, M. Hu, C-w Chang, R. L. Qiu, T. Liu, T. Liu, H. Mao, X. Yang Emory University, Atlanta, United States Impact: Our physics-informed deep learning model markedly reduces motion artifacts in MRI scans, enhancing image quality. By minimizing the need for repeat scans, this method could significantly decrease healthcare costs and bolster the reliability of downstream MR imaging applications.
		Computer Number: 32 3845. Robust Water-Fat Separation in Nasal MRI Using Hybrid Model and Data Driven Network View Presentation Video D. Li, K. Sun, J. Zhang, D. Shen School of Biomedical Engineering, & State Key Laboratory of Advanced Medical Materials and Devices, ShanghaiTech University, Shanghai, China Impact: Our proposed hybrid separation framework combines multi-task networks with a 3D physical model to enhance the robustness of water/fat separation in multi-contrast nasopharyngeal MRI images, significantly expanding its clinical applicability.

Similar Session(s)