Universal Generative Modeling for Calibration-free Parallel Mr Imaging
This addresses the need for faster and more efficient MRI acquisitions without requiring explicit coil calibration, though it appears incremental as it builds on existing deep learning and compressed sensing techniques.
The paper tackles the problem of calibration-free parallel MRI reconstruction by introducing an unsupervised deep learning framework called UGM-PI, which uses wavelet transforms and adaptive iteration to achieve results comparable or superior to state-of-the-art compressed sensing and parallel imaging methods.
The integration of compressed sensing and parallel imaging (CS-PI) provides a robust mechanism for accelerating MRI acquisitions. However, most such strategies require the explicit formation of either coil sensitivity profiles or a cross-coil correlation operator, and as a result reconstruction corresponds to solving a challenging bilinear optimization problem. In this work, we present an unsupervised deep learning framework for calibration-free parallel MRI, coined universal generative modeling for parallel imaging (UGM-PI). More precisely, we make use of the merits of both wavelet transform and the adaptive iteration strategy in a unified framework. We train a powerful noise conditional score network by forming wavelet tensor as the network input at the training phase. Experimental results on both physical phantom and in vivo datasets implied that the proposed method is comparable and even superior to state-of-the-art CS-PI reconstruction approaches.