Probabilistic self‐learning framework for low‐dose CT denoising

Bai, Ti; Wang, Biling; Nguyen, Dan; Jiang, Steve B

doi:10.1002/mp.14796

Cited by 29 publications

(33 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Tang et al [24] adopted CycleGAN [25] to train unpaired dataset for LDCT denoising. Although these methods have shown promising results, their performance is still inferior to that of supervised learning [26] and they have difficulty in preserving fine anatomical details due to their simple noise model [27].…”

Section: Plos Onementioning

confidence: 99%

Wavelet subband-specific learning for low-dose computed tomography denoising

et al. 2022

View full text Add to dashboard Cite

Deep neural networks have shown great improvements in low-dose computed tomography (CT) denoising. Early algorithms were primarily optimized to obtain an accurate image with low distortion between the denoised image and reference full-dose image at the cost of yielding an overly smoothed unrealistic CT image. Recent research has sought to preserve the fine details of denoised images with high perceptual quality, which has been accompanied by a decrease in objective quality due to a trade-off between perceptual quality and distortion. We pursue a network that can generate accurate and realistic CT images with high objective and perceptual quality within one network, achieving a better perception-distortion trade-off. To achieve this goal, we propose a stationary wavelet transform-assisted network employing the characteristics of high- and low-frequency domains of the wavelet transform and frequency subband-specific losses defined in the wavelet domain. We first introduce a stationary wavelet transform for the network training procedure. Then, we train the network using objective loss functions defined for high- and low-frequency domains to enhance the objective quality of the denoised CT image. With this network design, we train the network again after replacing the objective loss functions with perceptual loss functions in high- and low-frequency domains. As a result, we acquired denoised CT images with high perceptual quality using this strategy while minimizing the objective quality loss. We evaluated our algorithms on the phantom and clinical images, and the quantitative and qualitative results indicate that ours outperform the existing state-of-the-art algorithms in terms of objective and perceptual quality.

show abstract

Section: Plos Onementioning

confidence: 99%

Wavelet subband-specific learning for low-dose computed tomography denoising

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Studies [56]- [58] combined self-learning denoising methods with CT reconstruction to eliminate noise in reconstructed CT images. Bai et al [59] used probabilistic characteristics of denoising and Zhang et al [60] utilized adjacent CT slices. Zeng et al [61] proposed a LDCT sinogram recovery strategy based on unsupervised learning using the noise generation mechanism of CT measurement in the sinogram.…”

Section: Introductionmentioning

confidence: 99%

MM-Net: Multiframe and Multimask-Based Unsupervised Deep Denoising for Low-Dose Computed Tomography

Jeon

Kim

Choi

2023

IEEE Trans. Radiat. Plasma Med. Sci.

View full text Add to dashboard Cite

Low-dose computed tomography (LDCT) is crucial due to the risk of radiation exposure to patients. However, the high noise level in LDCT images may reduce the image quality, leading to a less accurate diagnosis. Deep learning technology, especially supervised methods, has recently been widely accepted as a powerful tool for LDCT image denoising tasks. However, supervised methods require numerous paired datasets of LDCT and high-quality pristine CT images, which are rarely available in real-world clinical scenarios. This study presents an unsupervised learning-based framework called MM-Net, consisting of two training steps for a volumetric LDCT denoising task. In the two-step training approach, we first train the initial denoising network multi-scale attention U-Net (MSAU-Net) in a self-supervised manner to predict the noise-suppressed center slice with a neighboring multi-slice input. The second training step aims to train the U-Net-based final denoiser based on the pre-trained MSAU-Net to improve the image quality by introducing new multi-patch and multi-mask matching loss. Qualitative visual inspection and quantitative measures across texturally different domains of clinical and animal data reveal that the proposed MM-Net outperformed all competing state-ofthe-art unsupervised algorithms. The unsupervised method also achieved denoising performance comparable to the representative supervised methods trained with ground truth images.

show abstract

“…Recently, deep learning approaches have become a main approach for lowdose CT reconstruction [17,4], which has resulted many commercially available products [9]. Furthermore, the difficulty of obtaining matched CT data pairs has led to exploring various unsupervised learning approaches [22,2]. In particular, image translation methodology is successfully employed for the low-dose CT noise suppression by learning the noise patterns through a comparison between the low-dose CT(LDCT) and high-dose CT(HDCT) distributions [20,7,16].…”

Section: Introductionmentioning

confidence: 99%

Patch-wise Deep Metric Learning for Unsupervised Low-Dose CT Denoising

Jung¹,

Lee²,

You³

et al. 2022

Preprint

View full text Add to dashboard Cite

The acquisition conditions for low-dose and high-dose CT images are usually different, so that the shifts in the CT numbers often occur. Accordingly, unsupervised deep learning-based approaches, which learn the target image distribution, often introduce CT number distortions and result in detrimental effects in diagnostic performance. To address this, here we propose a novel unsupervised learning approach for lowdose CT reconstruction using patch-wise deep metric learning. The key idea is to learn embedding space by pulling the positive pairs of image patches which shares the same anatomical structure, and pushing the negative pairs which have same noise level each other. Thereby, the network is trained to suppress the noise level, while retaining the original global CT number distributions even after the image translation. Experimental results confirm that our deep metric learning plays a critical role in producing high quality denoised images without CT number shift.

show abstract

Probabilistic self‐learning framework for low‐dose CT denoising

Cited by 29 publications

References 45 publications

Wavelet subband-specific learning for low-dose computed tomography denoising

Wavelet subband-specific learning for low-dose computed tomography denoising

MM-Net: Multiframe and Multimask-Based Unsupervised Deep Denoising for Low-Dose Computed Tomography

Patch-wise Deep Metric Learning for Unsupervised Low-Dose CT Denoising

Contact Info

Product

Resources

About