Saved in:
Bibliographic Details
Main Authors: Wu, Zihui, Yin, Tianwei, Sun, Yu, Frost, Robert, van der Kouwe, Andre, Dalca, Adrian V., Bouman, Katherine L.
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2304.12507
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917829721194496
author Wu, Zihui
Yin, Tianwei
Sun, Yu
Frost, Robert
van der Kouwe, Andre
Dalca, Adrian V.
Bouman, Katherine L.
author_facet Wu, Zihui
Yin, Tianwei
Sun, Yu
Frost, Robert
van der Kouwe, Andre
Dalca, Adrian V.
Bouman, Katherine L.
contents Compressed sensing magnetic resonance imaging (CS-MRI) seeks to recover visual information from subsampled measurements for diagnostic tasks. Traditional CS-MRI methods often separately address measurement subsampling, image reconstruction, and task prediction, resulting in a suboptimal end-to-end performance. In this work, we propose TACKLE as a unified co-design framework for jointly optimizing subsampling, reconstruction, and prediction strategies for the performance on downstream tasks. The naïve approach of simply appending a task prediction module and training with a task-specific loss leads to suboptimal downstream performance. Instead, we develop a training procedure where a backbone architecture is first trained for a generic pre-training task (image reconstruction in our case), and then fine-tuned for different downstream tasks with a prediction head. Experimental results on multiple public MRI datasets show that TACKLE achieves an improved performance on various tasks over traditional CS-MRI methods. We also demonstrate that TACKLE is robust to distribution shifts by showing that it generalizes to a new dataset we experimentally collected using different acquisition setups from the training data. Without additional fine-tuning, TACKLE leads to both numerical and visual improvements compared to existing baselines. We have further implemented a learned 4$\times$-accelerated sequence on a Siemens 3T MRI Skyra scanner. Compared to the fully-sampling scan that takes 335 seconds, our optimized sequence only takes 84 seconds, achieving a four-fold time reduction as desired, while maintaining high performance.
format Preprint
id arxiv_https___arxiv_org_abs_2304_12507
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Learning Task-Specific Strategies for Accelerated MRI
Wu, Zihui
Yin, Tianwei
Sun, Yu
Frost, Robert
van der Kouwe, Andre
Dalca, Adrian V.
Bouman, Katherine L.
Image and Video Processing
Computer Vision and Pattern Recognition
Compressed sensing magnetic resonance imaging (CS-MRI) seeks to recover visual information from subsampled measurements for diagnostic tasks. Traditional CS-MRI methods often separately address measurement subsampling, image reconstruction, and task prediction, resulting in a suboptimal end-to-end performance. In this work, we propose TACKLE as a unified co-design framework for jointly optimizing subsampling, reconstruction, and prediction strategies for the performance on downstream tasks. The naïve approach of simply appending a task prediction module and training with a task-specific loss leads to suboptimal downstream performance. Instead, we develop a training procedure where a backbone architecture is first trained for a generic pre-training task (image reconstruction in our case), and then fine-tuned for different downstream tasks with a prediction head. Experimental results on multiple public MRI datasets show that TACKLE achieves an improved performance on various tasks over traditional CS-MRI methods. We also demonstrate that TACKLE is robust to distribution shifts by showing that it generalizes to a new dataset we experimentally collected using different acquisition setups from the training data. Without additional fine-tuning, TACKLE leads to both numerical and visual improvements compared to existing baselines. We have further implemented a learned 4$\times$-accelerated sequence on a Siemens 3T MRI Skyra scanner. Compared to the fully-sampling scan that takes 335 seconds, our optimized sequence only takes 84 seconds, achieving a four-fold time reduction as desired, while maintaining high performance.
title Learning Task-Specific Strategies for Accelerated MRI
topic Image and Video Processing
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2304.12507