scholarly journals A novel strategy to develop deep learning for image super-resolution using original ultra-high-resolution computed tomography images of lung as training dataset

2021 ◽  
Author(s):  
Hitoshi Kitahara ◽  
Yukihiro Nagatani ◽  
Hideji Otani ◽  
Ryohei Nakayama ◽  
Yukako Kida ◽  
...  
Keyword(s):  
Computed Tomography ◽  
Deep Learning ◽  
High Resolution ◽  
Super Resolution ◽  
Training Dataset ◽  
High Resolution Computed Tomography ◽  
Computed Tomography Images ◽  
Image Super Resolution ◽  
Novel Strategy

scholarly journals Lung Segmentation on High-Resolution Computerized Tomography Images Using Deep Learning: A Preliminary Step for Radiomics Studies

2020 ◽  
Vol 6 (11) ◽  
pp. 125 ◽  
Author(s):  
Albert Comelli ◽  
Claudia Coronnello ◽  
Navdeep Dahiya ◽  
Viviana Benfante ◽  
Stefano Palmucci ◽  
...  
Keyword(s):  
Image Segmentation ◽  
Deep Learning ◽  
High Resolution ◽  
Idiopathic Pulmonary Fibrosis ◽  
Pulmonary Fibrosis ◽  
Dice Similarity Coefficient ◽  
Learning Models ◽  
High Resolution Computed Tomography ◽  
Computed Tomography Images ◽  
Self Driving Cars

Background: The aim of this work is to identify an automatic, accurate, and fast deep learning segmentation approach, applied to the parenchyma, using a very small dataset of high-resolution computed tomography images of patients with idiopathic pulmonary fibrosis. In this way, we aim to enhance the methodology performed by healthcare operators in radiomics studies where operator-independent segmentation methods must be used to correctly identify the target and, consequently, the texture-based prediction model. Methods: Two deep learning models were investigated: (i) U-Net, already used in many biomedical image segmentation tasks, and (ii) E-Net, used for image segmentation tasks in self-driving cars, where hardware availability is limited and accurate segmentation is critical for user safety. Our small image dataset is composed of 42 studies of patients with idiopathic pulmonary fibrosis, of which only 32 were used for the training phase. We compared the performance of the two models in terms of the similarity of their segmentation outcome with the gold standard and in terms of their resources’ requirements. Results: E-Net can be used to obtain accurate (dice similarity coefficient = 95.90%), fast (20.32 s), and clinically acceptable segmentation of the lung region. Conclusions: We demonstrated that deep learning models can be efficiently applied to rapidly segment and quantify the parenchyma of patients with pulmonary fibrosis, without any radiologist supervision, in order to produce user-independent results.

scholarly journals Identification of emphysema patterns in high resolution computed tomography images

2017 ◽  
Vol 4 (1) ◽  
pp. 16
Author(s):  
Musibau A. Ibrahim ◽  
Oladotun A. Ojo ◽  
Peter A. Oluwafisoye
Keyword(s):  
Computed Tomography ◽  
Fractal Dimension ◽  
High Resolution ◽  
Hölder Exponent ◽  
High Resolution Computed Tomography ◽  
Box Counting ◽  
Computed Tomography Images ◽  
Local Intensity ◽  
Box Counting Method

Fractal dimension (FD) is a very useful metric for the analysis of image structures with statistically self-similar properties. It has applications in areas such as texture segmentation, shape classification and analysis of medical images. Several approaches can be used for calculating the fractal dimension of digital images; the most popular method is the box-counting method. It is also very challenging and difficult to classify patterns in high resolution computed tomography images (HRCT) using this important descriptor. This paper applied the Holder exponent computation of the local intensity values for detecting the emphysema patterns in HRCT images. The absolute differences between the normal and the abnormal regions in the images are the key for a successful classification of emphysema patterns using the statistical analysis. The results obtained in this paper demonstrated the effectiveness of the predictive power of the features extracted from the Holder exponent in the analysis and classification of HRCT images. The overall classification accuracy achieved in lung tissue layers is greater than 90%, which is an evidence to prove the effectiveness of the methods investigated in this paper.

scholarly journals Paracoccidioidomycosis: level of pulmonary sequelae in high resolution computed tomography images from patients of two endemic regions of Brazil

2017 ◽  
Vol 7 (3) ◽  
pp. 318-325 ◽  
Author(s):  
Diana Rodrigues de Pina ◽  
Matheus Alvarez ◽  
Guilherme Giacomini ◽  
Ana Luiza Menegatti Pavan ◽  
Carlos Ivan Andrade Guedes ◽  
...  
Keyword(s):  
Computed Tomography ◽  
High Resolution ◽  
High Resolution Computed Tomography ◽  
Computed Tomography Images

scholarly journals Classification of High-Resolution Computed Tomography Manifestations for COVID-19 Patients: A Systematic Review

2021 ◽  
Vol In Press (In Press) ◽  
Author(s):  
Sanaz Alibabaei ◽  
Elham Rohollahpour ◽  
Marziyeh Tahmasbi
Keyword(s):  
Computed Tomography ◽  
High Resolution ◽  
False Negative ◽  
Disease Process ◽  
Common Finding ◽  
High Resolution Computed Tomography ◽  
Computed Tomography Images ◽  
Peripheral Areas ◽  
Severity Of The Disease ◽  
Low Sensitivity

Context: The early detection of COVID-19 is of paramount importance for the disease treatment and control. As real-time reverse-transcription polymerase chain reaction indicates a low sensitivity, the computed tomography of patients' chest can play an effective role in the diagnosis of COVID-19, particularly for patients with false-negative RT-PCR tests. It is also effective in monitoring the clinical trends and assessing the severity of the disease. Objectives: Accordingly, this study aimed to review the different manifestations of the COVID-19 infections in High-Resolution Computed Tomography images of patients' chests and analyze the distribution of the disease in the lungs. The results can contribute to providing a comprehensive and concise reference on the appearance of various types of involvement and lung lesions and the extent of these lesions in the COVID-19 patients. Data Sources: We systematically searched four major indexing databases (namely PubMed, Science Direct, Google Scholar, and Cochrane Central) for articles published by May 2021 using the following keywords: High-Resolution Computed Tomography (HRCT), COVID-19, and Manifestations. Results: Overall, 29 studies addressing the role of HRCT in detecting and evaluating the manifestations of the COVID-19 infection in patients' lungs as Ground Glass Opacification (GGO), Consolidation, Irregular Solid Nodules, Fibrous Stripes, Crazy Paving Pattern, Air Bronchogram Sign, etc. were reviewed. Conclusions: GGO was the most common finding, as reported in 96.6% of the reviewed articles, followed by Consolidations (65.5%) and Irregular Solid Nodules (55.2%). Most patients revealed the disease process as a bilateral distribution in the peripheral areas of the lung.

scholarly journals Reconstruction of High-Resolution Computed Tomography Image in Sinogram Space

2021 ◽  
Vol 15 ◽  
pp. 84-88
Author(s):  
Osama A. Omer
Keyword(s):  
Computed Tomography ◽  
High Resolution ◽  
Image Quality ◽  
Computational Cost ◽  
Super Resolution ◽  
Reconstruction Method ◽  
High Resolution Computed Tomography ◽  
Resolution Image ◽  
Reconstruction Methods

An important part of any computed tomography (CT) system is the reconstruction method, which transforms the measured data into images. Reconstruction methods for CT can be either analytical or iterative. The analytical methods can be exact, by exact projector inversion, or non-exact based on Back projection (BP). The BP methods are attractive because of thier simplicity and low computational cost. But they produce suboptimal images with respect to artifacts, resolution, and noise. This paper deals with improve of the image quality of BP by using super-resolution technique. Super-resolution can be beneficial in improving the image quality of many medical imaging systems without the need for significant hardware alternation. In this paper, we propose to reconstruct a high-resolution image from the measured signals in Sinogram space instead of reconstructing low-resolution images and then post-process these images to get higher resolution image.

Sign in / Sign up

Export Citation Format

Share Document