(IEEE Transactions on Computational Imaging)
1California Institute of Technology 2Massachusetts Institute of Technology 3Harvard Medical School
Compressed sensing magnetic resonance imaging (CS-MRI) seeks to recover visual information from subsampled measurements for diagnostic tasks. Traditional CS-MRI meth- ods 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×-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.
paper [arXiv][IEEE TCI] code [Github]
Zihui Wu, Tianwei Yin, Yu Sun, Robert Frost, Andre van der Kouwe, Adrian V. Dalca, and Katherine L. Bouman (2024). "Learning Task-Specific Strategies for Accelerated MRI." IEEE Transactions on Computational Imaging, vol. 10, pp. 1040-1054, 2024, doi: 10.1109/TCI.2024.3410521.
bibtex [show]
We propose a unified framework, called TAsk-specific Co-design of K-space subsampLing and prEdiction (TACKLE), for designing task-specific CS-MRI systems. We show that TACKLE enables better task performance than existing reconstruction-oriented methods on a variety of tasks summarized on the bottom left panel above.
In all the tasks we consider, TACKLE outperforms baselines both numerically and visually. We show some visual examples below.
TACKLE leads to substantial visual improvements on out-of-distribution data that we experimentally collect with different acquisition parameters.
A pair-wise comparison between TACKLE and baselines on each slice of the dataset demonstrates that improvement is both consistent and statistically significant.
