Neighbor2Neighbor: A Self-Supervised Framework for Deep Image Denoising

Asymmetric Convolution-based GAN Framework for Low-Dose CT Image Denoising

Noise reduction is essential to improve the diagnostic quality of low-dose CT (LDCT) images. In this regard, data-driven denoising methods based on generative adversarial networks (GAN) have shown promising results. However, custom designs with 2D convolution may not preserve the correlation of the local and global pixels, which results in the loss of high-frequency (edges/ boundaries of lesions) anatomical details. A recent state-of-the-art method demonstrates that using primitive GAN-based methods may introduce structural (shape) distortion. To address this issue, we develop a novel asymmetric convolution-based generator network (ACGNet), which is constructed by using one-dimensional (1D) asymmetric convolutions and a dynamic attention module (DAM). The 1D asymmetric convolutions (1 × 3 & 3 × 1) can intensify the representation power of square convolution kernels (3 × 3) in horizontal and vertical directions. Consequently, we integrated the highlighted low-level CT voxel details via purposed attention DAM with high-level CT-scan features. As a result, ACGNet efficiently preserves the local and global pixel relations in denoised LDCT images. Furthermore, we propose a novel neural structure preserving loss (NSPL) through which ACGNet learns the neighborhood structure of CT images, preventing structural (shape) distortion. In addition, the ACGNet can reconstruct the CT images with human-perceived quality via back-propagated gradients due to the feature-based NSPL loss. Finally, we include differential content loss in network optimization to restore high-frequency lesion boundaries. The proposed method outperforms many state-of-the-art methods on two publicly accessible datasets: the Mayo 2016 dataset (PSNR: 35.2015 dB, SSIM: 0.9560), and Low-dose CT image and projection dataset (PSNR: 35.2825 dB, SSIM: 0.9566).

Multi-Level Noise Sampling From Single Image for Low-Dose Tomography Reconstruction

Low-dose digital radiography (DR) and computed tomography (CT) become increasingly popular due to reduced radiation dose. However, they often result in degraded images with lower signal-to-noise ratios, creating an urgent need for effective denoising techniques. The recent advancement of the single-image-based denoising approach provides a promising solution without requirement of pairwise training data, which are scarce in medical imaging. These methods typically rely on sampling image pairs from a noisy image for inter-supervised denoising. Although enjoying simplicity, the generated image pairs are at the same noise level and only include partial information about the input images. This study argues that generating image pairs at different noise levels while fully using the information of the input image is preferable since it could provide richer multi-perspective clues to guide the denoising process. To this end, we present a novel Multi-Level Noise Sampling (MNS) method for low-dose tomography denoising. Specifically, MNS method generates multi-level noisy sub-images by partitioning the high-dimensional input space into multiple low-dimensional sub-spaces with a simple yet effective strategy. The superiority of the MNS method in single-image-based denoising over the competing methods has been investigated and verified theoretically. Moreover, to bridge the gap between self-supervised and supervised denoising networks, we introduce an optimization function that leverages prior knowledge of multi-level noisy sub-images to guide the training process. Through extensive quantitative and qualitative experiments conducted on large-scale clinical low-dose CT and DR datasets, we validate the effectiveness and superiority of our MNS approach over other state-of-the-art supervised and self-supervised methods.

Self-supervised learning for denoising of multidimensional MRI data

PURPOSE

To develop a fast denoising framework for high-dimensional MRI data based on a self-supervised learning scheme, which does not require ground truth clean image.

THEORY AND METHODS

Quantitative MRI faces limitations in SNR, because the variation of signal amplitude in a large set of images is the key mechanism for quantification. In addition, the complex non-linear signal models make the fitting process vulnerable to noise. To address these issues, we propose a fast deep-learning framework for denoising, which efficiently exploits the redundancy in multidimensional MRI data. A self-supervised model was designed to use only noisy images for training, bypassing the challenge of clean data paucity in clinical practice. For validation, we used two different datasets of simulated magnetization transfer contrast MR fingerprinting (MTC-MRF) dataset and in vivo DWI image dataset to show the generalizability.

RESULTS

The proposed method drastically improved denoising performance in the presence of mild-to-severe noise regardless of noise distributions compared to previous methods of the BM3D, tMPPCA, and Patch2self. The improvements were even pronounced in the following quantification results from the denoised images.

CONCLUSION

The proposed MD-S2S (Multidimensional-Self2Self) denoising technique could be further applied to various multi-dimensional MRI data and improve the quantification accuracy of tissue parameter maps.

Toward Ground-Truth Optical Coherence Tomography via Three-Dimensional Unsupervised Deep Learning Processing and Data

Optical coherence tomography (OCT) can perform non-invasive high-resolution three-dimensional (3D) imaging and has been widely used in biomedical fields, while it is inevitably affected by coherence speckle noise which degrades OCT imaging performance and restricts its applications. Here we present a novel speckle-free OCT imaging strategy, named toward-ground-truth OCT ( t GT-OCT), that utilizes unsupervised 3D deep-learning processing and leverages OCT 3D imaging features to achieve speckle-free OCT imaging. Specifically, our proposed t GT-OCT utilizes an unsupervised 3D-convolution deep-learning network trained using random 3D volumetric data to distinguish and separate speckle from real structures in 3D imaging volumetric space; moreover, t GT-OCT effectively further reduces speckle noise and reveals structures that would otherwise be obscured by speckle noise while preserving spatial resolution. Results derived from different samples demonstrated the high-quality speckle-free 3D imaging performance of t GT-OCT and its advancement beyond the previous state-of-the-art. The code is available online: https://github.com/Voluntino/tGT-OCT.

Unsupervised OCT image despeckling with ground-truth- and repeated-scanning-free features

Optical coherence tomography (OCT) can resolve biological three-dimensional tissue structures, but it is inevitably plagued by speckle noise that degrades image quality and obscures biological structure. Recently unsupervised deep learning methods are becoming more popular in OCT despeckling but they still have to use unpaired noisy-clean images or paired noisy-noisy images. To address the above problem, we propose what we believe to be a novel unsupervised deep learning method for OCT despeckling, termed Double-free Net, which eliminates the need for ground truth data and repeated scanning by sub-sampling noisy images and synthesizing noisier images. In comparison to existing unsupervised methods, Double-free Net obtains superior denoising performance when trained on datasets comprising retinal and human tissue images without clean images. The efficacy of Double-free Net in denoising holds significant promise for diagnostic applications in retinal pathologies and enhances the accuracy of retinal layer segmentation. Results demonstrate that Double-free Net outperforms state-of-the-art methods and exhibits strong convenience and adaptability across different OCT images.

An unsupervised two-step training framework for low-dose computed tomography denoising

BACKGROUND

Although low-dose computed tomography (CT) imaging has been more widely adopted in clinical practice to reduce radiation exposure to patients, the reconstructed CT images tend to have more noise, which impedes accurate diagnosis. Recently, deep neural networks using convolutional neural networks to reduce noise in the reconstructed low-dose CT images have shown considerable improvement. However, they need a large number of paired normal- and low-dose CT images to fully train the network via supervised learning methods.

PURPOSE

To propose an unsupervised two-step training framework for image denoising that uses low-dose CT images of one dataset and unpaired high-dose CT images from another dataset.

METHODS

Our proposed framework trains the denoising network in two steps. In the first training step, we train the network using 3D volumes of CT images and predict the center CT slice from them. This pre-trained network is used in the second training step to train the denoising network and is combined with the memory-efficient denoising generative adversarial network (DenoisingGAN), which further enhances both objective and perceptual quality.

RESULTS

The experimental results on phantom and clinical datasets show superior performance over the existing traditional machine learning and self-supervised deep learning methods, and the results are comparable to the fully supervised learning methods.

CONCLUSIONS

We proposed a new unsupervised learning framework for low-dose CT denoising, convincingly improving noisy CT images from both objective and perceptual quality perspectives. Because our denoising framework does not require physics-based noise models or system-dependent assumptions, our proposed method can be easily reproduced; consequently, it can also be generally applicable to various CT scanners or dose levels.

Pol2Pol: self-supervised polarimetric image denoising

In this Letter, we present a self-supervised method, polarization to polarization (Pol2Pol), for polarimetric image denoising with only one-shot noisy images. First, a polarization generator is proposed to generate training image pairs, which are synthesized from one-shot noisy images by exploiting polarization relationships. Second, the Pol2Pol method is extensible and compatible, and any network that performs well in supervised image denoising tasks can be deployed to Pol2Pol after proper modifications. Experimental results show Pol2Pol outperforms other self-supervised methods and achieves comparable performance to supervised methods.

MAS-Net OCT: a deep-learning-based speckle-free multiple aperture synthetic optical coherence tomography

High-resolution spectral domain optical coherence tomography (SD-OCT) is a vital clinical technique that suffers from the inherent compromise between transverse resolution and depth of focus (DOF). Meanwhile, speckle noise worsens OCT imaging resolving power and restricts potential resolution-enhancement techniques. Multiple aperture synthetic (MAS) OCT transmits light signals and records sample echoes along a synthetic aperture to extend DOF, acquired by time-encoding or optical path length encoding. In this work, a deep-learning-based multiple aperture synthetic OCT termed MAS-Net OCT, which integrated a speckle-free model based on self-supervised learning, was proposed. MAS-Net was trained on datasets generated by the MAS OCT system. Here we performed experiments on homemade microparticle samples and various biological tissues. Results demonstrated that the proposed MAS-Net OCT could effectively improve the transverse resolution in a large imaging depth as well as reduced most speckle noise.

Dual-scale similarity-guided cycle generative adversarial network for unsupervised low-dose CT denoising

Removing the noise in low-dose CT (LDCT) is crucial to improving the diagnostic quality. Previously, many supervised or unsupervised deep learning-based LDCT denoising algorithms have been proposed. Unsupervised LDCT denoising algorithms are more practical than supervised ones since they do not need paired samples. However, unsupervised LDCT denoising algorithms are rarely used clinically due to their unsatisfactory denoising ability. In unsupervised LDCT denoising, the lack of paired samples makes the direction of gradient descent full of uncertainty. On the contrary, paired samples used in supervised denoising allow the parameters of networks to have a clear direction of gradient descent. To bridge the gap in performance between unsupervised and supervised LDCT denoising, we propose dual-scale similarity-guided cycle generative adversarial network (DSC-GAN). DSC-GAN uses similarity-based pseudo-pairing to better accomplish unsupervised LDCT denoising. We design a Vision Transformer-based global similarity descriptor and a residual neural network-based local similarity descriptor for DSC-GAN to effectively describe the similarity between two samples. During training, pseudo-pairs, i.e., similar LDCT samples and normal-dose CT (NDCT) samples, dominate parameter updates. Thus, the training can achieve equivalent effect as training with paired samples. Experiments on two datasets demonstrate that DSC-GAN beats the state-of-the-art unsupervised algorithms and reaches a level close to supervised LDCT denoising algorithms.