Skip to content

lilygeorgescu/MHCA

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Multimodal Multi-Head Convolutional Attention with Various Kernel Sizes for Medical Image Super-Resolution (WACV 2023) - Official Repo

Mariana-Iuliana Georgescu, Radu Tudor Ionescu, Andreea-Iuliana Miron, Olivian Savencu, Nicolae-Catalin Ristea, Nicolae Verga and Fahad Shahbaz Khan.

🆕 This is the official repository of the "Multimodal Multi-Head Convolutional Attention with Various Kernel Sizes for Medical Image Super-Resolution" paper accepted at WACV 2023.

🌟 Overview

We propose a novel multimodal multi-head convolutional attention module for super-resolution. MHCA is a spatial-channel attention module that can be integrated into any neural network at any layer. We are also the first to perform medical image superresolution using a multimodal low-resolution input.

🔒 License

The present code is released under the Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) license.

💻 Code

We release the MHCA building block.

🚀 Results and trained models.

🌟 We obtained new SOTA results on T2w modality on the IXI data set for the scaling factor of 2x and 4x.

🤩 We release the pretrained models. Check EDSR folder and try out our models.

Method Scale PSNR/SSIM
EDSR + MCHA 2x 40.11/0.9871
EDSR + MMCHA 2x 40.28/0.9874
EDSR + MCHA 4x 32.15/0.9418
EDSR + MMCHA 4x 32.51/0.9452

🔨 Installation

Please follow the instructions in Install.md.

🖊Citation

Please cite our work if you use any material released in this repository.

@inproceedings{Georgescu-WACV-2023,
  title="{Multimodal Multi-Head Convolutional Attention with Various Kernel Sizes for Medical Image Super-Resolution}",
  author={Georgescu, Mariana-Iuliana and Ionescu, Radu Tudor and Miron, Andreea-Iuliana and Savencu, Olivian and Ristea, Nicolae-Catalin and Verga, Nicolae and Khan, Fahad Shahbaz},
  booktitle={Proceedings of WACV},
  year={2023},
  publisher={IEEE}
}

About

Multimodal Multi-Head Convolutional Attention with Various Kernel Sizes for Medical Image Super-Resolution

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published