Self-supervised dual-domain balanced dropblock-network for low-dose CT denoising

Objective.Self-supervised learning methods have been successfully applied for low-dose computed tomography (LDCT) denoising, with the advantage of not requiring labeled data. Conventional self-supervised methods operate only in the image domain, ignoring valuable priors in the sinogram domain. Recently proposed dual-domain methods address this limitation but encounter issues with blurring artifacts in the reconstructed image due to the inhomogeneous distribution of noise levels in low-dose sinograms.Approach.To tackle this challenge, this paper proposes SDBDNet, an end-to-end dual-domain self-supervised method for LDCT denoising. With the network designed based on the properties of inhomogeneous noise in low-dose sinograms and the principle of moderate sinogram-domain denoising, SDBDNet achieves effective denoising in dual domains without introducing blurring artifacts. Specifically, we split the sinogram into two subsets based on the positions of detector cells to generate paired training data with high similarity and independent noise. These sub-sinograms are then restored to their original size using 1D interpolation and learning-based correction. To achieve adaptive and moderate smoothing in the sinogram domain, we integrate Dropblock, a type of convolution layer with regularization, into SDBDNet, and set a weighted average between the denoised sinograms and their noisy counterparts, leading to a well-balanced dual-domain approach.Main results.Numerical experiments show that our method outperforms popular non-learning and self-supervised learning methods, demonstrating its effectiveness and superior performance.Significance.While introducing a novel high-performance dual-domain self-supervised LDCT denoising method, this paper also emphasizes and verifies the importance of appropriate sinogram-domain denoising in dual-domain methods, which might inspire future work.

Unpaired low-dose computed tomography image denoising using a progressive cyclical convolutional neural network

BACKGROUND

Reducing the radiation dose from computed tomography (CT) can significantly reduce the radiation risk to patients. However, low-dose CT (LDCT) suffers from severe and complex noise interference that affects subsequent diagnosis and analysis. Recently, deep learning-based methods have shown superior performance in LDCT image-denoising tasks. However, most methods require many normal-dose and low-dose CT image pairs, which are difficult to obtain in clinical applications. Unsupervised methods, on the other hand, are more general.

PURPOSE

Deep learning methods based on GAN networks have been widely used for unsupervised LDCT denoising, but the additional memory requirements of the model also hinder its further clinical application. To this end, we propose a simpler multi-stage denoising framework trained using unpaired data, the progressive cyclical convolutional neural network (PCCNN), which can remove the noise from CT images in latent space.

METHODS

Our proposed PCCNN introduces a noise transfer model that transfers noise from LDCT to normal-dose CT (NDCT), denoised CT images generated from unpaired CT images, and noisy CT images. The denoising framework also contains a progressive module that effectively removes noise through multi-stage wavelet transforms without sacrificing high-frequency components such as edges and details.

RESULTS

Compared with seven LDCT denoising algorithms, we perform a quantitative and qualitative evaluation of the experimental results and perform ablation experiments on each network module and loss function. On the AAPM dataset, compared with the contrasted unsupervised methods, our denoising framework has excellent denoising performance increasing the peak signal-to-noise ratio (PSNR) from 29.622 to 30.671, and the structural similarity index (SSIM) was increased from 0.8544 to 0.9199. The PCCNN denoising results were relatively optimal and statistically significant. In the qualitative result comparison, PCCNN without introducing additional blurring and artifacts, the resulting image has higher resolution and complete detail preservation, and the overall structural texture of the image is closer to NDCT. In visual assessments, PCCNN achieves a relatively balanced result in noise suppression, contrast retention, and lesion discrimination.

CONCLUSIONS

Extensive experimental validation shows that our scheme achieves reconstruction results comparable to supervised learning methods and has performed well in image quality and medical diagnostic acceptability.

CTformer: convolution-free Token2Token dilated vision transformer for low-dose CT denoising

Objective. Low-dose computed tomography (LDCT) denoising is an important problem in CT research. Compared to the normal dose CT, LDCT images are subjected to severe noise and artifacts. Recently in many studies, vision transformers have shown superior feature representation ability over the convolutional neural networks (CNNs). However, unlike CNNs, the potential of vision transformers in LDCT denoising was little explored so far. Our paper aims to further explore the power of transformer for the LDCT denoising problem.Approach. In this paper, we propose a Convolution-free Token2Token Dilated Vision Transformer (CTformer) for LDCT 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, overlapped inference mechanism is employed to effectively eliminate the boundary artifacts that are common for encoder-decoder-based denoising models.Main results. Experimental results on Mayo dataset suggest that the CTformer outperforms the state-of-the-art denoising methods with a low computational overhead.Significance. The proposed model delivers excellent denoising performance on LDCT. Moreover, low computational cost and interpretability make the CTformer promising for clinical applications.