Spach Transformer: Spatial and Channel-wise Transformer Based on Local and Global Self-attentions for PET Image Denoising

Se-In Jang, Tinsu Pan, Gary Y. Li, Pedram Heidari, Junyu Chen, Quanzheng Li, and Kuang Gong

News

• Nov 2023: Accepted in IEEE Transactions on Medical Imaging! [Paper]

Brief Introduction

• The focus of this project is on handling 3D PET input data.
• It incorporates a 3D-based approach, utilizing both the Swin Transformer and Restormer architectures specifically adapted for 3D data processing.

Installation

See INSTALL.md for the installation of dependencies required to run Spach Transformer. Do the following for a newer GPU (after activating your conda)

pip install --upgrade torch torchvision torchaudio

Quick Run

import torch
from model.SpachTransformer  import SpachTransformer
from model.Restormer         import Restormer

input   = torch.rand(1, 1, 96, 96, 96)
model1  = SpachTransformer()
output  = model1(input)

model2  = Restormer()
output  = model2(input)

---

Abstract: Position emission tomography (PET) is widely used in clinics and research due to its quantitative merits and high sensitivity, but suffers from low signal-to-noise ratio (SNR). Recently convolutional neural networks (CNNs) have been widely used to improve PET image quality. Though successful and efficient in local feature extraction, CNN cannot capture long-range dependencies well due to its limited receptive field. Global multi-head self-attention (MSA) is a popular approach to capture long-range information. However, the calculation of global MSA for 3D images has high computational costs. In this work, we proposed an efficient spatial and channel-wise encoder-decoder transformer, Spach Transformer, that can leverage spatial and channel information based on local and global MSAs. Experiments based on datasets of different PET tracers, i.e., 18F-FDG, 18F-ACBC, 18F-DCFPyL, and 68Ga-DOTATATE, were conducted to evaluate the proposed framework. Quantitative results show that the proposed Spach Transformer can achieve better performance than other reference methods.

---

Citation

If you use Spach Transformer, please consider citing:

@article{jang2022spach, 
    title={Spach Transformer: Spatial and channel-wise transformer based on local and global self-attentions for PET image denoising}, 
    author={Jang, Se-In and Pan, Tinsu and Li, Ye and Heidari, Pedram and Chen, Junyu and Li, Quanzheng and Gong, Kuang}, 
    journal={arXiv preprint arXiv:2209.03300}, 
    year={2022} }
}

Contact

Should you have any question, please contact [email protected]

Acknowledgment: This code is based on the Restormer and Swin Transformer.