Unsupervised MRI Reconstruction via Zero-Shot Learned Adversarial Transformers

TL;DR

SLATER is an unsupervised MRI reconstruction method that uses adversarial transformers to learn a high-quality MRI prior, enabling zero-shot reconstruction with superior performance compared to existing methods.

Contribution

Introduces SLATER, a novel zero-shot MRI reconstruction approach using adversarial transformers to learn priors without supervision, improving reconstruction quality.

Findings

SLATER outperforms state-of-the-art unsupervised methods on brain MRI datasets.

The method effectively captures long-range relationships in MRI images.

Zero-shot reconstruction achieves high data consistency and image quality.

Abstract

Supervised reconstruction models are characteristically trained on matched pairs of undersampled and fully-sampled data to capture an MRI prior, along with supervision regarding the imaging operator to enforce data consistency. To reduce supervision requirements, the recent deep image prior framework instead conjoins untrained MRI priors with the imaging operator during inference. Yet, canonical convolutional architectures are suboptimal in capturing long-range relationships, and priors based on randomly initialized networks may yield suboptimal performance. To address these limitations, here we introduce a novel unsupervised MRI reconstruction method based on zero-Shot Learned Adversarial TransformERs (SLATER). SLATER embodies a deep adversarial network with cross-attention transformers to map noise and latent variables onto coil-combined MR images. During pre-training, this unconditional network learns a high-quality MRI prior in an unsupervised generative modeling task. During inference, a zero-shot reconstruction is then performed by incorporating the imaging operator and optimizing the prior to maximize consistency to undersampled data. Comprehensive experiments on brain MRI datasets clearly demonstrate the superior performance of SLATER against state-of-the-art unsupervised methods.