Related papers: CTformer: Convolution-free Token2Token Dilated Vision Transformer for Low-dose CT Denoising

CTformer: Convolution-free Token2Token Dilated Vision Transformer for Low-dose CT Denoising

URL: http://arxiv.org/abs/2202.13517v1
Date: Mon, 28 Feb 2022 02:58:16 GMT
Title: CTformer: Convolution-free Token2Token Dilated Vision Transformer for Low-dose CT Denoising
Authors: Dayang Wang, Fenglei Fan, Zhan Wu, Rui Liu, Fei Wang, Hengyong Yu
Abstract summary: Low-dose computed tomography (LDCT) denoising is an important problem in CT research. vision transformers have shown superior feature representation ability over convolutional neural networks (CNNs) We propose a Convolution-free Token2Token Dilated Vision Transformer for low-dose CT denoising.
Score: 11.67382017798666
License: http://creativecommons.org/licenses/by/4.0/
Abstract: Low-dose computed tomography (LDCT) denoising is an important problem in CT research. Compared to the normal dose CT (NDCT), LDCT images are subjected to severe noise and artifacts. Recently in many studies, vision transformers have shown superior feature representation ability over convolutional neural networks (CNNs). However, unlike CNNs, the potential of vision transformers in LDCT denoising was little explored so far. To fill this gap, we propose a Convolution-free Token2Token Dilated Vision Transformer for low-dose CT denoising. The CTformer uses a more powerful token rearrangement to encompass local contextual information and thus avoids convolution. It also dilates and shifts feature maps to capture longer-range interaction. We interpret the CTformer by statically inspecting patterns of its internal attention maps and dynamically tracing the hierarchical attention flow with an explanatory graph. Furthermore, an overlapped inference mechanism is introduced to effectively eliminate the boundary artifacts that are common for encoder-decoder-based denoising models. Experimental results on Mayo LDCT dataset suggest that the CTformer outperforms the state-of-the-art denoising methods with a low computation overhead.