Artificial Intelligence based Covid19 detection using CTScan and Chest X-Ray Images: Machine Learning and Deep Learning techniques

Lung Disease is one of the most common healthcare issues in the entire world in all age groups of humans due to the cause of smoking, infection and contaminated air. Pneumonia, Coronavirus disease 2019 (COVID-19) and Tuberculosis are the common types of lung disease. To reduce the diagnosis of the Lung Disease, several research works have been done with the help of Artificial Intelligence based-technologies. Prediction and classification of Lung Disease are the common tasks using Chest X-ray images, CT scan images, and MRI images with the help of machine learning, deep learning, and transfer learning models. The main idea of this paper is to show the comparative analysis of ResNet-50, MobileNet-V2, VGG-19 and DenseNet201 CNN-based Transfer Learning techniques and suggesting which image classification algorithm is more recommended to predicting and classifying Lung Disease classification problem. Those four algorithms have been trained on the X-ray images and evaluated using performance metrics techniques. The results have been shown 94.64% accuracy obtained using ResNet50, 97.43% accuracy obtained using VGG19, 98.49% accuracy obtained using MobileNetV2, and 98.05% accuracy obtained using DenseNet201 pre-trained model. Based on the experimental results MobileNet model has been outperformed as compared to other models. Training time per epoch and ROC-AUC values of MobileNet was better than other models. Based on our studies, MobileNet transfer learning is recommended pre-trained model for Lung Disease prediction and classification problem.

PurposeTuberculosis (TB) is a severe bacterial infectious lung disease. Millions of people die or experience severe health complications due to TB each year. Accurate, automated, and effective detection of TB is key to curing and preventing person-to-person transmission. In this regard, deep learning (DL) and machine learning (ML) techniques applied to chest X-ray (CXR) images have proved effective. Here, we present our DL- and ML-based approach for TB detection using CXR images.Material and methodsWe implemented convolutional neural network (CNN)-based pre-trained DL models, such as DenseNet121, DenseNet169, DenseNet201, ResNet152, and VGG19, as feature extractors. ML models, including support vector machine (SVM), XGBoost, logistic regression, and a DL-based custom model, were used as classifiers. A total of 2,391 CXR images from three publicly available datasets were considered.ResultsWe found three models achieving the highest value in different evaluation metrics: accuracy (99.91%) with ResNet152 and SVM; recall (99.22%), precision (99.23%), and F1-score (99.22%) with DenseNet169 and the custom classifier; and area under the curve (99.99%) with DenseNet201 and the custom classifier. Among these models, we propose DenseNet169 with the custom classifier as the best performer for potential clinical application, as it reflected a high and well-balanced performance across all evaluation metrics.ConclusionsThis study evaluated pre-trained CNN-based DL and ML models on pre-processed CXR images for the detection of TB. The DenseNet169 model with a custom classifier stands out with its high and well-balanced performance, offering a significant contribution to the effective and automated detection of TB.

<title>Abstract</title> Bone cancer diagnosis is imperative for diagnosing and treating many forms of primary and metastatic bone cancers early. Traditional imaging techniques, such as CT, MRI, and X-ray scans, are effective but typically involve manual interpretation, which is laborious and prone to human mistake. With increased accuracy, dependability, and efficiency over conventional techniques, automated bone cancer diagnosis systems have been made possible by recent developments in machine learning (ML) and deep learning (DL). Even though a large portion of the recent literature has made significant strides in the diagnosis of bone cancer using deep learning techniques, many of these methods have numerous drawbacks, including computational complexity, overfitting in certain situations, and a lack of reliable databases. The objective of this research is to develop a method that can diagnose bone cancer as quickly, efficiently, and affordably as feasible. This method's main contribution is the combination of EfficientNetB4 and SVM algorithms, which improves accuracy, speed, and accessibility while making use of large datasets and reliable assessment measures. Combining EfficientNetB4 and SVM is crucial for diagnosing bone cancer because it harnesses the EfficientNetB4 technique's amazing capacity to extract useful features both quantitatively and qualitatively, as well as the SVM's significant advantage when it comes to binary separation. After a thorough analysis of numerous research, it was determined to combine these methods as they are distinguished by their great efficiency and simplicity in job implementation, especially in medical environments where accuracy and interpretability are of utmost importance The success of the suggested methods was shown by experiments on a large dataset (which contains 35244 X-ray images), which produced 98% precision, 97.47% recall, 98% accuracy, and 98% F1-score. The suggested method's better performance and computational economy are highlighted by comparison with machine learning, deep learning, and transfer learning techniques. Furthermore, the suggested system encounters a quick inference time of 41 ms, which qualifies it for clinical real-time applications. This study offers a viable strategy for early detection and better treatment outcomes by demonstrating the potential of integrating deep learning and traditional machine learning approaches for better bone cancer diagnosis.

Lung cancer disease prediction with CT scan and histopathological images feature analysis using deep learning techniques

Due to the COVID-19 Pandemic, doctors need to make medical decisions for their patients based on many examinations (e.g., polymerase chain reaction test, temperature test, CT-Scans, or X-rays). However, transfer learning has been used in several researches and focuses on only a single modality of biomarkers (e.g., CT-Scan or X-Ray) for diagnosing Pneumonia. In recent studies, a single modality has its own classification accuracy and every different biomarker may provide complementary information for detecting COVID-19 Pneumonia. The COVID-19 virus can be detected by CT-Scan and X-Ray imaging of the chest. In this work, we propose to use concatenation of two different transfer learning models using an open-source dataset of 2500 CT-Scan images and 2500 X-ray images for classifying CT-Scan images and X-ray images into two classes: normal and COVID-19 Pneumonia. We have used DenseNet121, MobileNet, Xception, InceptionV3, ResNet50, and VGG16 models for image recognition in our work. As a result, we achieve the best classification accuracy of 99.87% of the concatenation of ResNet50 and VGG16 networks. We also achieved the best classification accuracy of 98.00% when using a single modality of CT-Scan ResNet50 networks and classification accuracy of 98.93% for X-Ray VGG16 networks. Our multimodal fusion method shows a better classification accuracy compared to the method of using a single modality of biomarkers.

The World Health Organization recognizes pneumonia as a significant global health issue. Artificial intelligence, particularly machine learning, and deep learning has emerged as valuable tools for improving pneumonia diagnosis. However, these techniques face a major challenge: the lack of labeled data. To tackle this, we propose using unsupervised learning models, which can produce comparable results even with limited training data. Our study presents an unsupervised learning approach utilizing autoencoders to detect pneumonia from chest X-ray images. Our method uses Variational autoencoders for feature extraction, which are then employed in classification using a Random Forest classifier. The model is trained on a dataset containing two classes of X-ray images: pneumonia and normal. Our approach demonstrates effectiveness comparable to existing supervised learning methods.

One of the record killers in the world is lung disease. Lung disease denotes to many disorders affecting the lungs. These diseases can be identified through Chest X- Ray, Computed Tomography CT, Ultrasound tests. This study provides a systematic review on different types of Deep Learning (DL) designs, methods, techniques used by different researchers in diagnosing COVID-19, Pneumonia, Tuberculosis, Lung tumor, etc. In the present research study, a systematic review and analysis is carried by following PRISMA research methodology. For this study, more than 900 research articles are considered from various indexing sources such as Scopus and Web of Science. After several selection steps, finally a 40 quality research articles are included for detailed analysis. From this study, it is observed that majority of the research articles focused on DL techniques with Chest X-Ray images and few articles focused on CT scan images and very few have focused on Ultrasound images to identify the lung disease

The research community has recently shown significant interest in designing automated systems to detect coronavirus disease 2019 (COVID-19) using deep learning approaches and chest radiography images. However, state-of-the-art deep learning techniques, especially convolutional neural networks (CNNs), demand more learnable parameters and memory. Therefore, they may not be suitable for real-time diagnosis. Thus, the design of a lightweight CNN model for fast and accurate COVID-19 detection is an urgent need. In this paper, a lightweight CNN model called LW-CORONet is proposed that comprises a sequence of convolution, rectified linear unit (ReLU), and pooling layers followed by two fully connected layers. The proposed model facilitates extracting meaningful features from the chest X-ray (CXR) images with only five learnable layers. The proposed model is evaluated using two larger CXR datasets (Dataset-1: 2250 images and Dataset-2: 15,999 images) and the classification accuracy obtained are 98.67% and 99.00% on Dataset-1 and 95.67% and 96.25% on Dataset-2 for multi-class and binary classification cases, respectively. The results are compared with four contemporary pre-trained CNN models as well as state-of-the-art models. The effect of several hyperparameters: different optimization techniques, batch size, and learning rate have also been investigated. The proposed model demands fewer parameters and requires less memory space. Hence, it is effective for COVID-19 detection and can be utilized as a supplementary tool to assist radiologists in their diagnosis.

Medicine and engineering sciences have been working in close contact for common purposes. Machine learning algorithms are used in the medical field for early diagnosis prediction. The major aim of this study is to evaluate machine learning algorithms and deep learning algorithms using computed tomography scan (CT-scan) images for automated detection of the coronavirus disease 2019 (COVID-19) patients. We obtained seven hundred and fifty-seven (757) CT-scan images from a public platform. We applied four automated traditional classification methods to predict COVID-19 using deep learning and machine learning. These algorithms are SVM, AdaBoost, NASNetMobile, and InceptionV3. Comparative analyses are presented among the four models by considering metric performance factors to find the best model. The results show that the InceptionV3 model achieves better performance in terms of accuracy, precision, recall, Cohen’s kappa, F1- score, root mean squared error (RMSE), and receiver operating characteristic- area under the curve (ROC-AUC), in comparison with the other Covid-19 classifiers. Accordingly, the InceptionV3 approach is recommended for the automatic diagnosis of Covid-19 and assessments. This research can present a second point of view to medical experts and it can save time for researchers as the performance of standard machine learning methods in detecting COVID-19 is evaluated.

the SARS-COV2 (COVID-19) virus continues to spread over the world in the form of various mutant strains that are responsible for epidemic conditions. As a result, the clinical prognosis cannot be relied upon. Chest x-rays, computed tomography, and ultrasound imaging models supplement the analytical methods (for example, RT-PCR) to a certain degree, despite the fact that different clinical diagnostic approaches have been developed and used up to this point. The purpose of this work is to carry out the diagnosis of COVID-19 individuals utilizing the image padding technique, as well as to develop a Deep Covix-Net model for the identification of COVID-19, Pneumonia, and healthy individuals utilizing X-ray and CT scan images. Both of these objectives will be accomplished through this work. Using ultrasound pictures, create a model called Ultra Covix-Net for early diagnosis of COVID-19, pneumonia, and healthy persons; then, using CT scan images, estimate the severity of the COVID-19 infection. In order to complete the categorization process, machine learning and deep learning strategies are used.

COVID-19 is a contagious respiratory illness that can be passed from person to person. Because it affects the lungs, damages blood arteries, and causes cardiac problems, COVID-19 must be diagnosed quickly. The reverse transcriptase polymerase chain reaction (RT-PCR) is a method for detecting COVID-19, but it is time consuming and labor expensive, as well as putting the person collecting the sample in danger. As a result, clinicians prefer to use CT scan and Xray images. COVID-19 classification can be done manually, however AI makes the process go faster. AI approaches include image processing, machine learning, and deep learning. An AI model is required to diagnose COVID-19, and a dataset is necessary to train that model. A dataset consists of the information from which the model is trained. This paper consists of the review of different image processing, machine learning and deep learning papers proposed by different researchers. As well as models based on deep learning and pretrained model using gradient boosting algorithm The goal of this paper is to provide information for future researchers to work with.

Recent progress in COVID-19 detection techniques involve deep learning models. The patient’s image data like Chest X-Ray Images, CT-scan data help the physician for analyzing whether the patient is COVID-19 positive or negative. However, huge data size is essential for improving the classification accuracy of deep learning model. Data Augmentation (DA) is a promising solution to generate synthetic samples of data. Sampling is a traditional data augmentation technique to generate synthetic samples. Recently, Generative Adversarial Networks (GAN) have declared in generating high quality synthetic data from acutal small data to treat imbalance issue. This work proposed a method called GAN based Deep 2D-CNN (G-DCNN) for COVID-19 recognition. In this study, GAN has been used for synthesizing Chest X-Ray and CT-scan images followed by Deep 2D-CNN with the goal of detecting COVID-19.

Despite the vaccinations, the emergence of new and more contagious variants of the COVID-19 disease has continued to pose threats and challenges to our lives. Until herd immunity is achieved, it is important to continuously perform screening tests to control and minimize the transmissions. Due to the reported shortcomings of the RT-PCR, the utilization of deep learning for detecting COVID-19 from Chest X-Ray (CXR) images has gathered a lot of interest from researchers. As a contribution to the field, this study proposes a deep learning pipeline that utilizes transfer learning and image enhancement techniques to classify whether a given CXR image exhibits characteristics of COVID-19 infection, pneumonia infection, or normal/healthy lungs. For a lighter approach, the small pre-trained model named EfficientNetB0 is used as the base model for the transfer learning method. To improve the network’s performance, a sequence of contrast enhancement techniques namely the Multi-Scale Retinex (MSR) and Contrast Limited Adaptive Histogram Equalization (CLAHE) is introduced in the pipeline and employed as a pre-processing step. Gathered from a 10-fold cross-validation method in a dataset with 3729 images per class, results show that the proposed approach achieves an average overall accuracy of 92.089% with 98.431% average precision, 95.119% average recall, and 96.741% average f1-score for the COVID-19 class.

Chest X-ray imaging plays a vital role in the evaluation of thoracic organs, serving as a primary diagnostic tool for various pulmonary conditions. Interpretations of chest X-ray reports can sometimes differ from a physician’s clinical judgment, leading to diagnostic inconsistencies and potential delays in treatment. This study explores whether the application of advanced deep transfer learning techniques can improve the accuracy of chest X-ray interpretation. Deep Learning presents powerful capabilities for processing and interpreting complex imaging data. Convolutional Neural Networks (CNNs) are widely used in deep learning to perform image classification through hierarchical feature extraction. training deep models with large annotated datasets often demands significant computational resources. In this research, MobileNet and Inception V3 architectures are employed to detect diseases such as lung cancer, pneumonia, and tuberculosis from chest X-ray images. Conducted a detailed evaluation of these models and compare their performance with traditional diagnostic approaches. The results demonstrate notable improvements in accuracy and sensitivity, confirming that deep transfer learning techniques significantly enhance diagnostic outcomes. This suggests that MobileNet and Inception V3 can serve as reliable, efficient tools to support radiologists in early detection and decision-making, and contributing to more timely and accurate medical diagnoses.

Kidney stone detection is a critical healthcare challenge, as timely and accurate diagnosis can prevent complications. The motivation behind this review is the increasing prevalence of kidney stones and the need for more effective, non-invasive detection methods. Machine learning (ML) and deep learning (DL) techniques offer promising solutions, leveraging medical imaging data to enhance diagnostic accuracy. However, limitations such as high computational cost and reliance on large datasets hinder their full potential. The aim of this review is to analyze the latest advancements in kidney stone detection using ML and DL techniques. The objective is to compare existing methodologies, highlight their strengths and weaknesses, and suggest future research directions, particularly in integrating transfer learning and fine-tuning techniques to enhance performance.