Deep Transfer Learning-Based Framework for COVID-19 Diagnosis Using Chest CT Scans and Clinical Information

Abstract

The Coronavirus Disease 2019 (COVID-19) which first emerged in Wuhan, China in late December, 2019, has now spread to all the countries in the world. Conventional testing methods such as the antigen test, serology tests, and polymerase chain reaction tests are widely used. However, the test results can take anything from a few hours to a few days to reach the patient. Chest CT scan images have been used as alternatives for the detection of COVID-19 infection. Use of CT scan images alone might have limited capabilities, which calls attention to incorporating clinical features. In this paper, deep learning algorithms have been utilized to integrate the chest CT scan images obtained from patients with their clinical characteristics for fast and accurate diagnosis of COVID-19 patients. The framework uses an ANN to obtain the probability of the patient being infected with COVID-19 using their clinical information. Beyond a certain threshold, the chest CT scan of the patient is classified using a deep learning model which has been trained to classify the CT scan with 99% accuracy.

LDDNet: A Deep Learning Framework for the Diagnosis of Infectious Lung Diseases

“…In this study, the VGG-19 and DesNet-169 models were ensembled, and the authors claimed that this ensemble model showed better results than the other existing models for the case of five datasets, including three datasets of CT scan images and two datasets of X-ray images. Another study [ 17 ] proposed an ANN-based framework for the fast and automatic detection of patients infected by COVID-19. A study by Mukherjee et al [ 18 ] claimed that if multiple data types were integrated, then more information could be found, which might be helpful in detecting the anomaly patterns of COVID-19.…”

“…Secondly, RT-PCR tests are limited in availability and require considerable time. Deep learning (DL), a subset of AI, plays a dynamic role in controlling the outbreak of the virus infection, not only by detecting the presence of the virus during the early stages but also by enhancing the public health care system and analyzing the virus for appropriate medications and vaccination [ 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 ]. To identify the presence of abnormalities in the lung, DL can be used for the reconstruction and segmentation of chest X-rays or CT scans [ 13 , 14 ].…”

A Systematic Review on Deep Structured Learning for COVID-19 Screening Using Chest CT from 2020 to 2022

“…The use of pretrained DL models is prevalent in COVID-19 imaging tools. For example, commonly used models are AlexNet, VGGNet-XX, GoogleNet, ResNet-XX, Xception, Inception, wideResNet, MobileNet, NASNet, DarkNet, CheXNet, ShuffleNet, Incep-tionResNetV2, InceptionV3 (to name a few) [62,67,79,95,102,103,111,112,114,115,121,125,128,134,139,143,146,[151][152][153]158,159,162,210,211,213,214,220,234,247,249,255,263,270,274,277,284]. We found a total of 10 articles; among them, the researchers of 9 articles used conventional augmentation techniques such as resizing, zooming, gaussian noise, blur, spatial transformation, contrast adjustment, flipping, scaling, cropping, rotation, intensity transformation [69,71,75,84,85,92,93,96,…”

“…They reported that the proposed model AUC value is more than 80% on most test sides. Mishra et al [128] used a deep learning algorithm to diagnose COVID-19. The authors also worked on finding ANN's severity index of covid infection.…”

Imaging features and deep learning for prediction of pulmonary epithelioid hemangioendothelioma in CT images

“…Deep learning, a representative technique of artificial intelligence (AI), has shown great promise for detecting some common diseases based on clinical images, especially in medical image-related disease diagnosis and outcome prediction ( 11 ). Deep learning techniques have been utilized in image recognition of thoracic tumors ( 12 ).…”