Two stage residual CNN for texture denoising and structure enhancement on low dose CT image

Huang, Lei; Jiang, Huiyan; Shaojie, Li; Bai, Zhiqi; Zhang, Jitong

doi:10.1016/j.cmpb.2019.105115

Cited by 38 publications

(30 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…URL: https://imaging.nci.nih.gov/nbia-search-cover/ [ 3 , 4 , 6 , 13 , 20 , 45 , 50 , 79 , 91 ] D02 AAPM-Mayo Abdomen Mayo clinic AAPM Low- Dose CT Grand Challenge dataset consists of 2378 full and quarter dose CT images from 10 patients of 512 × 512 size. URL: https://www.aapm.org/GrandChallenge/LowDoseCT/ [ 3 , 6 , 12 , 13 , 16 , 23 , 30 , 34 , 36 , 42 , 46 , 58 , 71 , 73 , 75 , 77 , 78 , 81 , 83 , 91 ] D03 Piglet dataset [ 79 ] Whole-body Images were obtained under four dose levels and each dose level consists of 850 images of 512 × 512 size [ 20 , 75 , 79 ] D04 Data Science Bowl 2017 Lung The dataset consists of over a thousand high-resolution LDCT images of high-risk lung cancer patients. https://www.kaggle.com/c/data-science-bowl-2017/data [ 79 ] D05 ...…”

Section: Datasets and Methods To Deal With Data-related Issuesmentioning

confidence: 99%

“…https://www.kaggle.com/c/data-science-bowl-2017/data [ 79 ] D05 3D-IRCADb Different organs 1375 clinical NDCT images of 10 patients. URL: https://www.ircad.fr/research/3dircadb/ [ 30 ] D06 Luna-16 Lung 888 clinical NDCT scans are available with annotations. URL: https://luna16.grand-challenge.org/Data/ [ 82 ] Private datasets D06 Cardiac CT Cardiac Cardiac CT scans of 28 patients [ 72 ] D07 MGH dataset [ 76 ] Abdomen, chest, and head Massachusetts General Hospital (MGH) dataset consists of 40 cadaver scans obtained under four dose levels [ 58 ] D08 Cardiac CT Cardiac Two sets of 50 CT scans of mitral valve prolapse and coronary artery disease patients [ 35 ] D09 Dental CT Dental CT CT images were reconstructed using sinogram data in axial, sagittal, and coronal planes [ 28 ] D10 Liver simulated dataset Liver and portal vein 2480 NDCT images of liver and portal vein of 62 patients [ 53 ] D11 Brain clinical dataset Brain 200 ...…”

Section: Datasets and Methods To Deal With Data-related Issuesmentioning

confidence: 99%

“…However, the Siddon ray-driven algorithm and radon transformation-based algorithms simulate the Poisson noise into the low-dose transmission data as a product of simulated low-dose scan incident flux and the exponential of inverse sinogram. The studies [ 20 , 30 ] have used the radon transform-based algorithm, whereas [ 91 ] has used the forward projection-based algorithm for LDCT reconstruction.…”

Section: Datasets and Methods To Deal With Data-related Issuesmentioning

confidence: 99%

“…Estimating the residuals at each sub-band by the ResNet and implementation of concatenation later in the network are the specific features in Wave-ResNet in contrast to AAPM-Net. Apart from that, the two-stage denoising model (TS- RCNN2) in [ 30 ] has been trained using the stationary wavelet transformed LDCT and averaged-NDCT images. The two ResNets in this application have performed texture preservation and structure enhancement, respectively.…”

Section: Architecturesmentioning

confidence: 99%

“…In addition to this, if an LDCT image is transformed into the frequency domain, the noise content of the LDCT image will distribute as the high frequencies in LDCT images. Thus, it can be observed that some studies applied wavelet transformation to LDCT images for estimating and removing these noise-induced frequencies iteratively [ 30 , 36 , 71 , 75 ]. After the noise frequency filtering, the residual low-frequency information in the LDCT images can process through the DL model.…”

Section: Functional Aspectsmentioning

confidence: 99%

See 4 more Smart Citations

A review on Deep Learning approaches for low-dose Computed Tomography restoration

Kulathilake

Abdullah

Sabri

et al. 2021

Complex Intell. Syst.

View full text Add to dashboard Cite

Computed Tomography (CT) is a widely use medical image modality in clinical medicine, because it produces excellent visualizations of fine structural details of the human body. In clinical procedures, it is desirable to acquire CT scans by minimizing the X-ray flux to prevent patients from being exposed to high radiation. However, these Low-Dose CT (LDCT) scanning protocols compromise the signal-to-noise ratio of the CT images because of noise and artifacts over the image space. Thus, various restoration methods have been published over the past 3 decades to produce high-quality CT images from these LDCT images. More recently, as opposed to conventional LDCT restoration methods, Deep Learning (DL)-based LDCT restoration approaches have been rather common due to their characteristics of being data-driven, high-performance, and fast execution. Thus, this study aims to elaborate on the role of DL techniques in LDCT restoration and critically review the applications of DL-based approaches for LDCT restoration. To achieve this aim, different aspects of DL-based LDCT restoration applications were analyzed. These include DL architectures, performance gains, functional requirements, and the diversity of objective functions. The outcome of the study highlights the existing limitations and future directions for DL-based LDCT restoration. To the best of our knowledge, there have been no previous reviews, which specifically address this topic.

show abstract

Section: Datasets and Methods To Deal With Data-related Issuesmentioning

confidence: 99%

Section: Datasets and Methods To Deal With Data-related Issuesmentioning

confidence: 99%

Section: Datasets and Methods To Deal With Data-related Issuesmentioning

confidence: 99%

Section: Architecturesmentioning

confidence: 99%

Section: Functional Aspectsmentioning

confidence: 99%

See 3 more Smart Citations

A review on Deep Learning approaches for low-dose Computed Tomography restoration

Kulathilake

Abdullah

Sabri

et al. 2021

Complex Intell. Syst.

View full text Add to dashboard Cite

show abstract

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

Sadia,

Chen,

Zhang

2024

J Applied Clin Med Phys

View full text Add to dashboard Cite

With the ever‐increasing use of computed tomography (CT), concerns about its radiation dose have become a significant public issue. To address the need for radiation dose reduction, CT denoising methods have been widely investigated and applied in low‐dose CT images. Numerous noise reduction algorithms have emerged, such as iterative reconstruction and most recently, deep learning (DL)‐based approaches. Given the rapid advancements in Artificial Intelligence techniques, we recognize the need for a comprehensive review that emphasizes the most recently developed methods. Hence, we have performed a thorough analysis of existing literature to provide such a review. Beyond directly comparing the performance, we focus on pivotal aspects, including model training, validation, testing, generalizability, vulnerability, and evaluation methods. This review is expected to raise awareness of the various facets involved in CT image denoising and the specific challenges in developing DL‐based models.

show abstract

A CNN‐based denoising method trained with images acquired with electron density phantoms for thin‐sliced coronary artery calcium scans

Yang,

Hou

2024

J Applied Clin Med Phys

View full text Add to dashboard Cite

PurposeThis work proposed a convolutional neural network (CNN)‐based method trained with images acquired with electron density phantoms to reduce quantum noise for coronary artery calcium (CAC) scans reconstructed with slice thickness less than 3 mm.MethodsA DenseNet model was used to estimate quantum noise for CAC scans reconstructed with slice thickness of 0.5, 1.0 and 1.5 mm. Training data was acquired using electron density phantoms in three different sizes. The label images of the CNN model were real noise maps, while the input images of the CNN model were pseudo noise maps. Image denoising was conducted by subtracting the CNN output images from thin‐sliced CAC scans. The efficacy of the proposed method was verified through both phantom study and patient study.ResultsBy means of phantom study, the proposed method was proven effective in reducing quantum noise in CAC scans reconstructed with 1.5‐mm slice thickness without causing significant texture change or variation in HU values. With regard to patient study, calcifications were more clear on the denoised CAC scans reconstructed with slice thickness of 0.5, 1.0 and 1.5 mm than on 3‐mm slice images, while over‐smooth changes were not observed in the denoised CAC scans reconstructed with 1.5‐mm slice thickness.ConclusionOur results demonstrated that the electron density phantoms can be used to generate training data for the proposed CNN‐based denoising method to reduce quantum noise for CAC scans reconstructed with 1.5‐mm slice thickness. Because anthropomorphic phantom is not a necessity, our method could make image denoising more practical in routine clinical practice.

show abstract

Two stage residual CNN for texture denoising and structure enhancement on low dose CT image

Cited by 38 publications

References 26 publications

A review on Deep Learning approaches for low-dose Computed Tomography restoration

A review on Deep Learning approaches for low-dose Computed Tomography restoration

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

A CNN‐based denoising method trained with images acquired with electron density phantoms for thin‐sliced coronary artery calcium scans

Contact Info

Product

Resources

About