A mask-guided attention deep learning model for COVID-19 diagnosis based on an integrated CT scan images database

The global extent of COVID-19 mutations and the consequent depletion of hospital resources highlighted the necessity of effective computer-assisted medical diagnosis. COVID-19 detection mediated by deep learning models can help diagnose this highly contagious disease and lower infectivity and mortality rates. Computed tomography (CT) is the preferred imaging modality for building automatic COVID-19 screening and diagnosis models. It is well-known that the training set size significantly impacts the performance and generalization of deep learning models. However, accessing a large dataset of CT scan images from an emerging disease like COVID-19 is challenging. Therefore, data efficiency becomes a significant factor in choosing a learning model. To this end, we present a multi-task learning approach, namely, a mask-guided attention (MGA) classifier, to improve the generalization and data efficiency of COVID-19 classification on lung CT scan images. The novelty of this method is compensating for the scarcity of data by employing more supervision with lesion masks, increasing the sensitivity of the model to COVID-19 manifestations, and helping both generalization and classification performance. Our proposed model achieves better overall performance than the single-task (without MGA module) baseline and state-of-the-art models, as measured by various popular metrics.

Classic Augmentation based Classifier Generative Adversarial Network (CACGAN) for COVID-19 Diagnosis

Computer-aided diagnosis of COVID-19 from lung medical images has received increasing attention in previous clinical practice and research. However, developing such automatic model is usually challenging due to the requirement of a large amount of data and sufficient computer power. With only 317 training images, this paper presents a Classic Augmentation based Classifier Generative Adversarial Network (CACGAN) for data synthetising. In order to take into account, the feature extraction ability and lightness of the model for lung CT images, the CACGAN network is mainly constructed by convolution blocks. During the training process, each iteration will update the discriminator's network parameters twice and the generator's network parameters once. For the evaluation of CACGAN. This paper organized multiple comparison between each pair from CACGAN synthetic data, classic augmented data, and original data. In this paper, 7 classifiers are built, ranging from simple to complex, and are trained for the three sets of data respectively. To control the variable, the three sets of data use the exact same classifier structure and the exact same validation dataset. The result shows the CACGAN successfully learned how to synthesize new lung CT images with specific labels. © 2023 SPIE.

Fine-Tuning ResNet-152V2 for COVID-19 Recognition in Chest Computed Tomography Images

The epidemic Covid-19 has extended to majority of nations. This pandemic is due to a contagious condition 'SARS-CoV-2', was identified by the the International Health association. In order to diagnosis this virus from 2D chest computed tomography (CT) images, we applied three different transfer learning algorithms: $VGG-19, ResNet-152V2$ and a Fine-Tuned version of $ResNet-152V2$. The different transfer learning models are used on three hundred and four exams where 74 are normal cases, 60 are community-acquired pneumonia (CAP) cases and 169 were confirmed corona-virus cases. The best accuracy value is reached by the fine-tuned $ResNet-152v2$ by 75% against 70% for the basic $ResNet-152v2$ and 66% for the $VGG-19$. © 2022 IEEE.

Identification and Localization of COVID-19Abnormalities on Chest Radiographs using Computer Vision

The Corona Virus (COVID 19) pandemic is quickly becoming the world's most deadly disease. The spreading rate is higher and the early detection helps in faster recovery. The existence of COVID 19 in individuals shall be detected using molecular analysis or through radiographs of lungs. As time and test kit are limited RT- PCR is not suitable to test all. The RT- PCR being a time-consuming process, diagnosis using chest radiographs needs no transportation as the modern X-ray systems are digitized. Deep learning takes an edge over other techniques as it deduces the features automatically and performs massively parallel computations. Multiple feature maps will help in accurate prediction. The objective of the proposed work is to develop a Computer Aided Deep Learning System identify and localize COVID-19 virus from other viruses and pneumonia. It helps to detect COVID-19 within a short period of time thereby improving the lifetime of the individuals. SIIM-FISABIO-RSNA benchmark datasets are used to examine the proposed system. Recall, Precision, Accuracy-rate, and F-Measure are the metrics used to prove the integrity of the system. © 2023 IEEE.

Towards Understanding the COVID-19 Specific Patterns in Lung CT Scan Images using Deep CTNET Framework

Automated diagnosis of COVID-19 based on CTScan images of the lungs has caught maximum attention by many researchers in recent times. The rationale of this work is to exploit the texture patterns viz. deep learning networks so that it reduces the intra-class similarities among the patterns of COVID-19, Pneumonia and healthy class samples. The challenge of understanding the concurrence of the patterns of COVID-19 with other closely related patterns of other lung diseases is a new challenge. In this paper, a fine-tuned variational deep learning architecture named Deep CT-NET for COVID-19 diagnosis is proposed. Variation modelling to Deep CT-NET is evaluated using Resnet50, Xception, InceptionV3 and VGG19. Initially, grey level texture features are exploited to understand the correlation characteristics between these grey level patterns of COVID-19, Pneumonia and Healthy class samples. CT scan image dataset of 20,978 images was used for experimental analysis to assess the performance of Deep CT-NET viz., all mentioned models. Evaluation outcomes reveals that Resnet50, Xception, and InceptionV3 producing better performance with testing accuracy more than 96% in comparison with VGG19. © 2022 IEEE.

Classification of Lung CT Scan Images using Machine Learning

Timely discovery of COVID-19 may safeguard numerous diseased people. Several such lung diseases can turn to be life threatening. Early detection of these diseases can help in treating them at an early stage before it becomes threatening. In this paper, the proposed 3D CNN model helps in classifying the CT scans as normal and abnormal, which can then be used to treat the patients after recognizing the diseases. Chest X-ray is fewer commanding in the initial phases of the sickness, while a CT scan of the chest is advantageous even formerly symptoms seem, and CT scan accurately identify the anomalous features which are recognized in images. Besides this, using the two forms of images will raise the database. This will enhance the classification accuracy. In this paper the model used is a 3D CNN model;using this model the predictions are done. The dataset used is acquired from NKP Salve Medical Institute, Nagpur. This acquired dataset is used for prediction while an open source database is used for training the CNN model. After training the model the prediction were successfully completed, with these proposed 3D CNN model total accuracy of 87.86% is achieved. This accuracy can further be increased by using larger dataset. © 2022 IEEE.

Mathematical Morphology Adopted Automatic Detection of Ground Glass Opacities in Lung CT Images of COVID-19 Patients

The severity of the nCOVID infection relies on the presence of Ground Glass Opacities (GGO) present in the patient's chest CT scan images. Although, detecting and delineating the precise boundaries of GGO in the chest CT images is challenging. Here, we proposed a fast and novel technique to automatically segment the regions containing GGO in lung CT images using mathematical morphology. We have tested our algorithm on the chest CT images of 145 Covid-positive cases. This unique segmentation approach correctly segments the lung field from chest CT images and identifies GGO with average sensitivity, specificity, and accuracy of 96.89%, 95.23%, and 97.22%, respectively. We used expert radiologists' hand-curated segmentation of GGO as ground truth for quantificational performance analysis. Our research results indicate that this algorithm performs well found in the literature. © 2023 IEEE.

COVID-19 Identification From Lung CT Scans in a Low-Resource Setting Using a Regularized 3D Convolutional Neural Network

Right at the end of 2019, the world saw an outbreak of a new type of SARS (severe acute respiratory syndrome) disease, SARS-Cov-2, or COVID-19. Even in 2022, around 1 million people worldwide are getting infected with the virus every day. To date, more than 6 million people have died as a result of the virus. To tackle the pandemic, the first step is to successfully detect the virus among the mass population. The most popular method is the RT-PCR test, which, unfortunately, is not always conclusive. The physicians thus suggest lung CT tests for the patients for clinical relevance. But the problem with lung CT scans for the detection of coronavirus is that the COVID-19 infected scan is very similar to community-affected pneumonia (CAP) infected scan, and the results in many cases get wrongly interpreted. In addition, the virus is always mutating into different strains, and the severity and infection pattern slightly change with each mutation. Because of this rapid mutation, a large and balanced dataset of lung CT scans is not always available. In this work, we systematically evaluate the accuracy of a deep 3D convolutional neural network (CNN) on a small-scale and highly imbalanced dataset of lung CT scans (the SPGC COVID 2021 dataset). Our experiments show that it can outperform previous state-of-the-art 3D CNN models with proper regularization, an appropriate number of dense layers, and a weighted loss function. Our research, therefore, suggests an effective solution for identifying COVID-19 in lung CT scans using deep learning for small and highly imbalanced datasets. © 2022 IEEE.

Deep Learning based Lung Segmentation Prior for Robust COVID-19 Classification

Independent of a person's race, COVID-19 is one of the most contagious diseases in the world. The World Health Organization classified the COVID-19 outbreak as a pandemic after noting its global distribution. By using (i) sample-supported analysis and (ii) image-assisted diagnosis, COVID-19 is examined and verified. Our goal is to use CT scan images to identify the COVID-19 infiltrates. The followings steps are used to carry out the suggested work: (i) Automated segmentation with CNN;(ii) Feature mining;(iii) Principal feature selection with Bat-Algorithm;(iv) Classifier implementation using mobile framework and (v) Performance evaluation. We used a variety of automatic segmentation algorithms in our experiment, and the VGG-16 produced better results. This study is evaluated using benchmark datasets gathered, and SVM based RBF kernal classifier system resulted in superior COVID-19 abnormality identification. © 2023 IEEE.

Universal Adversarial Perturbation Attack on the Inception-Resnet-v1 model and the Effectiveness of Adversarial Retraining as a Suitable Defense Mechanism

In this study, we analyse the impact of the Universal Adversarial Perturbation Attack on the Inception-ResNet-v1 model using the lung CT scan dataset for COVID-19 classification and the retinal OCT scan dataset for Diabetic Macular Edema (DME) classification. The effectiveness of adversarial retraining as a suitable defense mechanism against this attack is examined. This study is categorised into three sections-The implementation of the Inception-ResNet-v1 model, the effect of the attack and the adversarial retraining. © 2023 IEEE.

Detection Covid-19 Infection of Lung CT Scan Slices Images Based on a Transfer Learning and GRAD-CAM

The long timeline of polymerase chain reaction (PCR) tests and the lack of test tool kits in many hospitals lead to fast infection according to the slow diagnosis. The Various experiences of radiologists cause deferent in accurately detection lessons. This research suggested and designed a model based on utilizing the deep learning (DL) algorithms to detect and visualize the infection of covid-19 patients. This work shows how various convolution layers of convolution neural networks (CNN) extract different types of features, which helps us understand how CNN steadily gains spatial information in each layer. As a result, every transition focuses on the region of interest. Understanding how CNN identifies and locates distinct infection areas in an image is easier with a heat map of activation. A gradient class activation map (Grad-CAM) was used to visualize the infected area in the lungs. Transfer learning such as Resnet50, VGG16, and inception V3 have been applied to the dataset to detect the infection area. The result of these models reached an accuracy of 71.01%, 83.51%, and 94.93%, respectively. The VGG16 has been manipulated because the model consumes less training time than the other models to solve the problem. Manipulating on VGG16 has been accomplished to achieve acceptable accuracy. The tuning on the last three layers of VGG16 architecture (dense layers) replaces them with two layers (global average layer and dense layer). The dense layer that is added deals with binary classification problems depending on the sigmoid function. This tuning serves the current study by speeding up the model's prediction and increasing the accuracy. The result of the testing reached 0.9683% of accuracy and 0.0931 loss function without augmentation. This work depends on the current dataset because it contains the mask of lung and infection, which was obtained to localize the infection area. The obtained result proved that the system could help the radiologist accommodate the pandemic. © 2022 IEEE.

Data analysis of COVID-19 pandemic using CT images

The COVID-19, infectious disease which is caused by the new virus called corona virus [SARS-COV-2], majorly affects the lungs which can be identified by the CT scan of the corona affected patients. We have collected for about 353 lung CT frames from each COVID-19 patient. it is about 369 non COVID-19 CT frames for the purpose of testing and training which gives the best identification technique in this pandemic situation. The identification technique we have introduced in this paper is data augmentation technique which gave best results that will be discussed here in further. From the collected data, we have utilized 75% of the lungs CT frames for training and another 25% frames for difficult attributes. This research paper encompasses of the results specified on CT images of corona complete, improved and loss problems which helps in comparative analysis. So, this comparative analysis of CT images is the illustration investigative statistics examination.

Hyperparameter Search for CT-Scan Classification Using Hyperparameter Tuning in Pre-Trained Model CNN With MLP

The common approach to find best hyperparameter in CNN training is grid search, by observing one set to another of hyperparameter for obtaining the best result. However, this approach is considered inefficient, time-consuming, and ineffectively computational. In this study, we are observing 2 hyperparameter tuning algorithms (bayesian optimization and random search) in search of the best hyperparameter for CT-Scan classification case. The used dataset is COVID-19 and non-COVID-19 lung CT-Scans. Several CNN architectures are also used such as: InceptionV4, MobileNetV3, and EfficientnetV2 with additional multi-layer perceptron on top layers. Based on the experiments, model EfficientnetV2-L architecture using hyperparameter from bayesian-optimization can outperform other models, with batch size of 32, learning rate of 0.01, dropout 0.5, Adam optimizer and SoftMax activation, resulting in the accuracy rate of 0.94% and a model training time of 50 minutes 40 seconds. © 2022 IEEE.

PDRF-Net: a progressive dense residual fusion network for COVID-19 lung CT image segmentation

The lungs of patients with COVID-19 exhibit distinctive lesion features in chest CT images. Fast and accurate segmentation of lesion sites from CT images of patients' lungs is significant for the diagnosis and monitoring of COVID-19 patients. To this end, we propose a progressive dense residual fusion network named PDRF-Net for COVID-19 lung CT segmentation. Dense skip connections are introduced to capture multi-level contextual information and compensate for the feature loss problem in network delivery. The efficient aggregated residual module is designed for the encoding-decoding structure, which combines a visual transformer and the residual block to enable the network to extract richer and minute-detail features from CT images. Furthermore, we introduce a bilateral channel pixel weighted module to progressively fuse the feature maps obtained from multiple branches. The proposed PDRF-Net obtains good segmentation results on two COVID-19 datasets. Its segmentation performance is superior to baseline by 11.6% and 11.1%, and outperforming other comparative mainstream methods. Thus, PDRF-Net serves as an easy-to-train, high-performance deep learning model that can realize effective segmentation of the COVID-19 lung CT images. © 2023, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.

Explainability of deep learning models in medical image classification

The ability to explain the reasons for one's decisions to others is an important aspect of being human intelligence. We will look at the explainability aspects of the deep learning models, which are most frequently used in medical image processing tasks. The Explainability of machine learning models in medicine is essential for understanding how the particular ML model works and how it solves the problems it was designed for. The work presented in this paper focuses on the classification of lung CT scans for the detection of COVID-19 patients. We used CNN and DenseNet models for the classification and explored the application of selected visual explainability techniques to provide insight into how the model works when processing the images. © 2022 IEEE.

COVID Detection and Severity Prediction with 3D-ConvNeXt and Custom Pretrainings

Since COVID strongly affects the respiratory system, lung CT-scans can be used for the analysis of a patients health. We introduce a neural network for the prediction of the severity of lung damage and the detection of a COVID-infection using three-dimensional CT-data. Therefore, we adapt the recent ConvNeXt model to process three-dimensional data. Furthermore, we design and analyze different pretraining methods specifically designed to improve the models ability to handle three-dimensional CT-data. We rank 2nd in the 1st COVID19 Severity Detection Challenge and 3rd in the 2nd COVID19 Detection Challenge. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Transfer learning-based CNN diagnostic framework for diagnosis of COVID-19 from lung CT images

COVID-19 is an infectious disease that affects the respiratory system. To assist the physician in diagnosing lung disorders from chest CT images various systems have been developed and used. Detection of COVID-19 remains a challenging area of research. The objective of the work is to develop an inductive parameter-transfer learning-based approach for the prediction of COVID-19, pneumonia, from lung CT images. Our proposed approach is built on layer wise and convolution block-wise fine-tuning which designs the CNN architecture highly specific to lung CT image. We implemented the DenseNet201, InceptionV3, Xception, VGG19, and ResNet50 as baseline models. The network architectures are developed to learn feature representation of lung CT images. For the experimental analysis, five datasets are used. From the experimental results, it is inferred that the DenseNet201 model yields higher accuracy of 0.94 for Adam optimizer and 0.93 for the RMSprop optimizer compared to other models. © 2023 The Royal Photographic Society.

A Novel 18-Convolutional Layered Deep U-Net Architecture for COVID-19 Infection Diagnosis Through Object Detection on Lung CT Scan Segmentation

With the advancement of digital technology, a large number of medical images are created in the field of digital medicine. Object detection is a critical field of study in medical image analysis, detection and processing. Object detection is the categorization of the pixels in an image that determines where the objects in the image are located. Image classification is the global classification of an image's pixels. The objects are identified but not located from all of the pixels in the image. The intelligent identification process for the adjuvant diagnosis to assist medical doctors of various disciplines is in high demand. The need for the high accurate object detection method for the medical images remains a major challenge toward disease detection in the health care. With this summary overview, a novel model is proposed for lung CT scan image detection as 18-Convolutional Layered Deep U-Net architecture for diagnosis of COVID-19 detection through object detection. Model fitting is done with 18-Convolutional Layered Deep U-Net architecture, and the model design is formed with four contraction path and four expansive path layers. The customized model design is fine-tuned with the parameter optimization. The object detection is done, and the performance is analyzed. The dataset is also fitted with existing deep convolution neural network models like region-based, threshold, edge-based and clustering-based method, and the performance is analyzed and compared with proposed model with metrics like pixel accuracy, intersection over union and dice coefficient. Implementation results show that the proposed models have pixel accuracy of 98.32%, intersection over union of 48.7% and dice coefficient of 97.56% compared to existing object detection models. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Two-Stage COVID19 Classification Using BERT Features

We propose an automatic COVID1-19 diagnosis framework from lung CT-scan slice images using double BERT feature extraction. In the first BERT feature extraction, A 3D-CNN is first used to extract CNN internal feature maps. Instead of using the global average pooling, a late BERT temporal pooing is used to aggregate the temporal information in these feature maps, followed by a classification layer. This 3D-CNN-BERT classification network is first trained on sampled fixed number of slice images from every original CT scan volume. In the second stage, the 3D-CNN-BERT embedding features are extracted for every 32 slice images sequentially, and these features are divided into fixed number of segments. Then another BERT network is used to aggregate these features into a single feature followed by another classification layer. The classification results of both stages are combined to generate final outputs. On the validation dataset, we achieve macro F1 score 92.05%;and on the testing dataset, we achieve macro F1 84.43%. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Lung CT image segmentation algorithm based on improved U-net structure

Accurate segmentation of medical images can help doctors diagnose and treat diseases. In the face of the complex COVID-19 image, this paper proposes an improved U-net network segmentation model, which uses the residual network structure to deepen the network level, and adds the attention module to integrate different receptive field, global, local and spatial features to enhance the detail segmentation effect of the network. For the COVID-19 CT data set, the F1-Score, Accuracy, SE, SP and Precision of the U-Net network are 0.9176, 0.9578, 0.9669, 0.9487 and 0.8574 respectively. Compared with U-Net, our model proposed in this paper increased by 6.43%, 3.36%, 0.85%, 4.78% and 13.11% on F1-Score, Accuracy, SE, SP and Precision, respectively. The automatic and effective segmentation of COVID-19 lung CT image is realized. © 2022 IEEE.