MR Image Super Resolution By Combining Feature Disentanglement CNNs and Vision Transformers

Dwarikanath Mahapatra, Zongyuan Ge

Keywords: MRI, super resolution, disentanglement, CNN, ViT

TL;DR: MR image super-resolution method that combines the advantages of CNNs and Vision transformers.

Abstract: State of the art magnetic resonance (MR) image super-resolution methods (ISR) leverage limited contextual information due to the use of CNNs, which learn interactions over a small neighborhood. On the other hand Vision transformers (ViT) have the ability to learn much more global contextual information, which is especially relevant for MR ISR since they provide additional information to generate superior quality HR images. We propose to combine local information of CNNs and global information from ViTs for image super resolution and output super resolved images that have superior quality than those produced by state of the art methods. Additionally, we incorporate extra constraints through multiple novel loss functions that preserve structure and texture information from the low resolution to high resolution images.

Registration: I acknowledge that publication of this at MIDL and in the proceedings requires at least one of the authors to register and present the work during the conference.

Authorship: I confirm that I am the author of this work and that it has not been submitted to another publication before.

Paper Type: both

Primary Subject Area: Application: Other

Secondary Subject Area: Image Acquisition and Reconstruction

Confidentiality And Author Instructions: I read the call for papers and author instructions. I acknowledge that exceeding the page limit and/or altering the latex template can result in desk rejection.

Code And Data:

1. The data that was used for the paper is publicly available

2. We are unable to share the code due to intellectual property rules of our company, which restricts making public certain research that can be commercialized.