SuperFormer: Volumetric Transformer Architectures for MRI Super-Resolution

TL;DR

SuperFormer introduces a volumetric transformer architecture tailored for MRI super-resolution, effectively capturing 3D anatomical details and outperforming traditional 3D CNN methods on the Human Connectome Project dataset.

Contribution

The paper extends Swin Transformers to 3D medical imaging and proposes a novel volumetric transformer framework for MRI super-resolution.

Findings

SuperFormer outperforms 3D CNN-based methods.

Effective encoding of 3D positional information.

Leverages multi-domain information for better reconstruction.

Abstract

This paper presents a novel framework for processing volumetric medical information using Visual Transformers (ViTs). First, We extend the state-of-the-art Swin Transformer model to the 3D medical domain. Second, we propose a new approach for processing volumetric information and encoding position in ViTs for 3D applications. We instantiate the proposed framework and present SuperFormer, a volumetric transformer-based approach for Magnetic Resonance Imaging (MRI) Super-Resolution. Our method leverages the 3D information of the MRI domain and uses a local self-attention mechanism with a 3D relative positional encoding to recover anatomical details. In addition, our approach takes advantage of multi-domain information from volume and feature domains and fuses them to reconstruct the High-Resolution MRI. We perform an extensive validation on the Human Connectome Project dataset and demonstrate the superiority of volumetric transformers over 3D CNN-based methods. Our code and pretrained models are available at https://github.com/BCV-Uniandes/SuperFormer.