ISMRM25 - Deep Learning for Image Enhancement: Part I

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

Deep Learning for Image Enhancement: Part I

Digital Poster

AI & Machine Learning

Wednesday, 14 May 2025

Exhibition Hall

15:45 - 16:45

Session Number: D-28

No CME/CE Credit

		Computer Number: 1 3814. OptiFlow: Deep Learning-Based Motion Estimation and Frame Interpolation for Brain MRI View Presentation Video H. Neeli, I. Seimenis, D. Martin, N. Tsekos, P. Martin University of Houston, Houston, United States Impact: This work serves as a proof of concept, in which our optical flow-based deep learning technique has the potential to reduce MRI acquisition times, overcome issues of image distortions, and enhance overall image quality.
		Computer Number: 2 3815. Spatially Adaptive DWI Denoising using Plug-and-Play Diffusion Models View Presentation Video M. Mostapha, R. Miron, N. Janardhanan, M. Nadar, O. Darwish, T. Huelnhagen, T. Würfl, D. Grodzki, R. Schneider Siemens Healthineers, Princeton, United States Impact: We present a denoising method that accelerates DWI scans through a PnP diffusion model that utilizes noise maps for guidance. This approach improves scanning efficiency while preserving image quality, showcasing promise for future DWI clinical applications.
		Computer Number: 3 3816. Physics-encoded Neural Network for the correction of encoding magnetic field in a low-field gradient-coil-free permanent magnet MRI system View Presentation Video H. Jing Han, W. Yu, S. Huang Singapore University of Technology and Design, Singapore, Singapore Impact: PeNN effectively and efficiently corrects the spatial encoding magnetic field in a portable MRI system using model-based imaging, saving regular labor-intensive mapping of magnetic fields. The proposed PeNN does not necessarily require the physics concept to be differentiable for backpropagation.
		Computer Number: 4 3817. A Registration-Based Framework for Generating Aligned CT Images from Misaligned CBCT-CT Pairs View Presentation Video A. Shazly, D. Kim, A. Al-Fakih, A. Rezk, K. Ryu, M. Al-masni Sejong University, Seoul, Korea, Republic of Impact: This methodology eliminates the need for perfectly aligned inputs, which are often unavailable in practice. By using the misaligned CT image as a proxy label, the proposed self-supervised approach leverages CBCT information to enhance the high-quality aligned CT (aCT) format.
		Computer Number: 5 3818. Denoising MRSI Data Using Subspace Model-Assisted Langevin Dynamics with Side Information View Presentation Video W. Jin, Z. Xu, Y. Li, Y. Zhao, R. Guo, Y. Li, Z-P Liang Beckman Institute for Advanced Science and Technology, University of Illinois at Urbana-Champaign, Urbana, United States Impact: This proposed method provides an effective way to enhance the sensitivity of MRSI using physics-based machine learning incorporating side information. The method may further enhance the practical utility of MRSI for research and clinical applications.
		Computer Number: 6 3819. 3D consistent data restoration for corrupted MRI slices using DDPMs with posterior sampling View Presentation Video S. Vasylechko, O. Afacan, S. Kurugol Boston Children's Hospital, Harvard Medical School, Boston, United States Impact: The demonstrated combination of diffusion posterior sampling with self-supervised learning establishes a framework for artifact-robust medical image restoration. This advances both computational efficiency in MRI post-processing and enables new research into automated quality assessment of cardiac functional metrics.
		Computer Number: 7 3820. Coordinate-Based Neural Representation for Motion-Robust 3D Multiparametric Quantitative MRI with Fat Navigators View Presentation Video G. Lao, X. Zong, Y. Zhang, H. Wei Shanghai Jiao Tong University, Shanghai, China Impact: The proposed method can simultaneously generate motion-robust multiparametric quantitative maps of the whole brain without the need for k-space correction, increasing the clinical usability of multiparametric quantitative MRI.
		Computer Number: 8 3821. Nex2Nex – An unsupervised deep learning method for denoising MR images View Presentation Video S. Chatterjee, R. Sundaresan, S. Rajamani, D. Shanbhag GE HealthCare, Bengaluru, India Impact: Proposed method enables training MRI DL denoising models without requirement to accurately model MRI noise. This approach becomes helpful while training denoising methods for low SNR MRI data.
		Computer Number: 9 3822. Deep Learning MR Image Denoising with Synthetic Phase and Noise View Presentation Video Z. Zhou, L. Xiang, H. Gandhi, A. Shankaranarayanan Subtle Medical Inc, Menlo Park, United States Impact: Synthetic phase allows convenient clinical deployment of complex DL denoising (CpxDN) models that shows advantage in high level and structured noise suppression. More clinical evaluation and optimization on CpxDN performance worth further investigation.
		Computer Number: 10 3823. Unsupervised Multi-Contrast MRI Super-Resolution with the Implicit Feature Sampling and Reciprocal Framework W. Chen, Y. Fan, Z. Li, C. Liu, Y. Wang, Q. Tian, D. Shen, X. Song Tsinghua University, Beijing, China Impact: Our model facilitates multi-contrast MRI super-resolution in the absence of ground-truth HR images, which not only substantially reduces MRI acquisition time, but enables the obtaining of certain HR sequences that are difficult to acquire in clinical settings.
		Computer Number: 11 3824. A model-based deep learning (DL) approach for denoising MR images View Presentation Video S. Chatterjee, M. Lebel GE HealthCare, Bengaluru, India Impact: A model-based approach for MRI denoising provides guarantees against contrast and fine structure alterations during denoising. By virtue of the proposed method, we shall have robust and generalizable MRI denoisers.
		Computer Number: 12 3825. Self-Supervised Denoising Model for 7T MR Angiography with Z-Directional Artery Focused DDPM View Presentation Video S. Yun, S-H You, B. Kim, D-H Kim, H. Seo Bionics Research Center, Korea Institute of Science and Technology, Seoul, Korea, Republic of Impact: This study advances 7T MRA imaging by providing a ground truth-free denoising method, potentially improving the diagnosis accuracy of cerebrovascular diseases and enhancing the clinical utility of 7T MRI in vascular imaging applications.
		Computer Number: 13 3826. Spatial-Angular Representation Learning for High-Fidelity Super-Resolution in Diffusion MRI View Presentation Video R. Wu, J. Cheng, C. Li, J. Zou, W. Fan, Y. Liang, S. Wang Paul C. Lauterbur Research Center for Biomedical Imaging, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen 518055, China Impact: Simultaneously enhancing the spatial and angular resolution of dMRI data can effectively reduce acquisition time, improve the accuracy of quantitative parameter estimation, and enhance clinical diagnostic efficiency.
		Computer Number: 14 3827. Reconstruction of Super-Resolution T1 Maps Using Efficient Residual Denoising Diffusion Probabilistic Models at 7T Field View Presentation Video M. Safari, Z. Eidex, S. Wang, C-W Chang, R. L. Liu, H. Mao, E. H. Middlebrooks, X. Yang Emory University, Atlanta, United States Impact: The proposed model can reduce the scan time required for generating high-resolution T1 maps within a clinically acceptable time. Its capacity to produce high-quality brain images with reduced artifacts may improve diagnosis and accelerate advancements in neuroimaging research.
		Computer Number: 15 3828. 1H-Guided Unsupervised Super-Resolution Reconstruction of 23Na in Simulcast X-nuclei MRI P. Cheng, C. Yang, C. Wu, K. Wang, L. Yang, L. Hu, J. Yuan, X. Sun NHC and CAMS Key Laboratory of Molecular Probe and Targeted Theranostics, Molecular Imaging Research Center (MIRC) of Harbin Medical University, Harbin, China Impact: Our deep learning approach markedly improves non-proton nuclei image quality in simulcast multi-nuclei MRI, advancing rigorous scientific research in simulcast multi-nuclei MRI, especially in settings with limited access to advanced multi-nuclei simultaneous imaging technologies.
		Computer Number: 16 3829. Robust domain adaptation for transferable MRI denoising models View Presentation Video C. Zaki, C. Millard, M. Chiew University of Toronto, Toronto, Canada Impact: This research demonstrates improved denoising performance on low-noise OOD MRI data, addressing a key challenge in generalizing to diverse imaging conditions with limited training data.

Similar Session(s)