Transfer learning framework for low‐dose CT reconstruction based on marginal distribution adaptation in multiscale

Abstract

Background: With the increasing use of computed tomography (CT) in clinical practice, limiting CT radiation exposure to reduce potential cancer risks has become one of the important directions of medical imaging research. As the dose decreases, the reconstructed CT image will be severely degraded by projection noise. Purpose: As an important method of image processing, supervised deep learning has been widely used in the restoration of low-dose CT (LDCT) in recent years. However, the normal-dose CT (NDCT) corresponding to a specific LDCT (it is regarded as the label of the LDCT, which is necessary for supervised learning) is very difficult to obtain so that the application of supervised learning methods in LDCT reconstruction is limited. It is necessary to construct a unsupervised deep learning framework for LDCT reconstruction that does not depend on paired LDCT-NDCT datasets. Methods: We presented an unsupervised learning framework for the transferring from the identity mapping to the low-dose reconstruction task, called marginal distribution adaptation in multiscale (MDAM). For NDCTs as source domain data, MDAM is an identity map with two parts: firstly, it establishes a dimensionality reduction mapping, which can obtain the same feature distribution from NDCTs and LDCTs; and then NDCTs is retrieved by reconstructing the image overview and details from the low-dimensional features. For the purpose of the feature transfer between source domain and target domain (LDCTs), we introduce the multiscale feature extraction in the MDAM, and then eliminate differences in probability distributions of these multiscale features between NDCTs and LDCTs through wavelet decomposition and domain adaptation learning. Results: Image quality evaluation metrics and subjective quality scores show that, as an unsupervised method, the performance of the MDAM approaches or even surpasses some state-of -the-art supervised methods. Especially, MDAM has been favorably evaluated in terms of noise suppression, structural preservation, and lesion detection. Conclusions: We demonstrated that, the MDAM framework can reconstruct corresponding NDCTs from LDCTs with high accuracy, and without relying on any labeles. Moreover, it is more suitable for clinical application compared with supervised learning methods.

CT image denoising methods for image quality improvement and radiation dose reduction

“… 150 , 151 with the remaining approaches including filter‐based methods as well as hybrid methods. 152 , 153 , 154 , 155 , 156 , 157 To be noted that some models are originally developed, 61 , 62 , 65 , 66 , 68 , 70 , 72 , 73 , 77 , 122 , 153 while some are developed by modifying original models through modifying loss functions, or layers, or extending original models to different domains. 60 , 63 , 64 , 69 , 71 , 72 , 74 , 75 , 76 , 78 , 82 , 84 , 102 , 121 , 123 , 124 , 125 , 147 , 150 , 151 , 152 , 155 , 158 …”

“…Instead, they use techniques such as GANs, VAEs, or self‐supervised learning to estimate the underlying distribution of clean images from the noisy images. Twenty studies 85 , 87 , 120 , 121 , 122 , 127 , 128 , 129 , 130 , 131 , 133 , 135 , 136 , 141 , 143 , 145 , 146 , 148 , 149 , 154 , 158 apply different unsupervised training approaches. Unsupervised DL‐based methods rely on the assumption that the noisy image can be modeled as a combination of a clean image and additive noise, and aim to estimate the clean image from the noisy input.…”