Figure - available from: Soft Computing
This content is subject to copyright. Terms and conditions apply.
Source publication
COVID-19 is a virus that causes upper respiratory tract and lung infections. The number of cases and deaths increased daily during the pandemic. Once it is vital to diagnose such a disease in a timely manner, the researchers have focused on computer-aided diagnosis systems. Chest X-rays have helped monitor various lung diseases consisting COVID-19....
Similar publications
The use of masks has become crucial in combating the Coronavirus pandemic. Unfortunately, the regulation of wearing a mask is not being upheld by many citizens which is contributing to the spread of the disease. To aid the efforts of regulations and to maintain safety in public areas, both large like parks or small like public transport, Artificial...
Citations
... Medical imaging such as X-ray and CT scans have been widely used to detect the coronavirus. Authors in [23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38]and [39] have used x-ray image data to predict and diagnose COVID-19. ...
... The research conducted by Kaya & Gürsoy [38] used a 3616-sample dataset for detecting covid-19 using X-ray images. They proposed a novel fine-tuning mechanism for COVID-19 infection detection and applied it to a deep transfer learning model. ...
... The researchers in [31,33,35,36,38,43,47], and [49] utilize both images and symptoms data, indicating their emphasis on diagnosing COVID-19 after the onset of symptoms. This is evident from their reliance on symptom data and X-ray images obtained post-confirmation of the disease. ...
With the outbreak of COVID-19 in 2020, countries worldwide faced significant concerns and challenges. Various studies have emerged utilizing Artificial Intelligence (AI) and Data Science techniques for disease detection. Although COVID-19 cases have declined, there are still cases and deaths around the world. Therefore, early detection of COVID-19 before the onset of symptoms has become crucial in reducing its extensive impact. Fortunately, wearable devices such as smartwatches have proven to be valuable sources of physiological data, including Heart Rate (HR) and sleep quality, enabling the detection of inflammatory diseases. In this study, we utilize an already-existing dataset that includes individual step counts and heart rate data to predict the probability of COVID-19 infection before the onset of symptoms. We train three main model architectures: the Gradient Boosting classifier (GB), CatBoost trees, and TabNet classifier to analyze the physiological data and compare their respective performances. We also add an interpretability layer to our best-performing model, which clarifies prediction results and allows a detailed assessment of effectiveness. Moreover, we created a private dataset by gathering physiological data from Fitbit devices to guarantee reliability and avoid bias.
The identical set of models was then applied to this private dataset using the same pre-trained models, and the results were documented. Using the CatBoost tree-based method, our best-performing model outperformed previous studies with an accuracy rate of 85% on the publicly available dataset. Furthermore, this identical pre-trained CatBoost model produced an accuracy of 81% when applied to the private dataset. You will find the source code in the link: https://github.com/OpenUAE-LAB/Covid-19-detection-using-Wearable-data.git.
... Medical imaging such as X-ray and CT scans have been widely used to detect the coronavirus. Authors in [23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38]and [39] have used x-ray image data to predict and diagnose COVID-19. ...
... The research conducted by Kaya & Gürsoy [38] used a 3616-sample dataset for detecting covid-19 using X-ray images. They proposed a novel fine-tuning mechanism for COVID-19 infection detection and applied it to a deep transfer learning model. ...
... The researchers in [31,33,35,36,38,43,47], and [49] utilize both images and symptoms data, indicating their emphasis on diagnosing COVID-19 after the onset of symptoms. This is evident from their reliance on symptom data and X-ray images obtained post-confirmation of the disease. ...
With the outbreak of COVID-19 in 2020, countries worldwide faced significant concerns and challenges. Various studies have emerged utilizing Artificial Intelligence (AI) and Data Science techniques for disease detection. Although COVID-19 cases have declined, there are still cases and deaths around the world. Therefore, early detection of COVID-19 before the onset of symptoms has become crucial in reducing its extensive impact. Fortunately, wearable devices such as smartwatches have proven to be valuable sources of physiological data, including Heart Rate (HR) and sleep quality, enabling the detection of inflammatory diseases. In this study, we utilize an already-existing dataset that includes individual step counts and heart rate data to predict the probability of COVID-19 infection before the onset of symptoms. We train three main model architectures: the Gradient Boosting classifier (GB), CatBoost trees, and TabNet classifier to analyze the physiological data and compare their respective performances. We also add an interpretability layer to our best-performing model, which clarifies prediction results and allows a detailed assessment of effectiveness. Moreover, we created a private dataset by gathering physiological data from Fitbit devices to guarantee reliability and avoid bias.
The identical set of models was then applied to this private dataset using the same pre-trained models, and the results were documented. Using the CatBoost tree-based method, our best-performing model outperformed previous studies with an accuracy rate of 85% on the publicly available dataset. Furthermore, this identical pre-trained CatBoost model produced an accuracy of 81% when applied to the private dataset. You will find the source code in the link: https://github.com/OpenUAE-LAB/Covid-19-detection-using-Wearable-data.git.
... In [18], the authors have used modified MobileNet for diagnosis of COVID-19. The authors of [19] proposed a deep transfer learning approach with fine tuning for the classification of COVID-19 patients using Chest X-ray images. In [20], the authors have experimented transfer learning approaches using various deep learning models for COVID-19 detection from Chest X-ray and CT scan images, and also proposed a modification of ResNet50 architecture. ...
The pandemic of COVID-19 has affected worldwide population. Diagnosing this highly contagious disease at an initial stage is essential for controlling its spread. In this paper, we propose a novel lightweight hybrid convolutional neural network to identify COVID-19-positive patients from Chest X-ray images using building blocks of MobileNet and ResNet50. This model shows improvement over the base models when applied to the same dataset independently. The proposed model is trained and tested on same configuration and dataset as that of its base models, for better performance analysis and comparison. It also has fewer parameters as compared to MobileNet and ResNet50, which makes it lightweight, adding to its better performance. It achieves an accuracy of 98.47 % with very few false negative predicted samples, which is extremely beneficial in medical diagnosis. Concerning the parameters, this model is built with only 200,034 trainable parameters, which marks a parameter reduction percentage of 99.15 % concerning ResNet50 and 93.82% compared to MobileNet. The generalisation performance of our proposed model is evaluated using stratified 5-fold cross validation technique. Our model achieved very good accuracy with very less parameters as compared to various state-of-the-art models. The model also classified Pneumonia images from Normal ones with an accuracy of 98.9 %.
... To benchmark our model's efficacy, we compared it against a range of established classification models under identical experimental conditions. These models included non-residual networks such as AlexNet and ZFNet; residual networks like ResNet-18, ResNet-34, ResNet-50, ResNet-101, and ResNet-152; and lightweight networks such as MobileNet [22], SqueezeNet [23], and ShuffleNet [24]. Comparative evaluation focused on metrics such as accuracy (ACC), precision, recall, F1-score, the number of model parameters, and inference time, enabling a comprehensive analysis of model performance across different architectures. ...
The accurate selection of the ultrasound plane for the fetal head and pubic symphysis is critical for precisely measuring the angle of progression. The traditional method depends heavily on sonographers manually selecting the imaging plane. This process is not only time-intensive and laborious but also prone to variability based on the clinicians' expertise. Consequently, there is a significant need for an automated method driven by artificial intelligence. To enhance the efficiency and accuracy of identifying the pubic symphysis-fetal head standard plane (PSFHSP), we proposed a streamlined neural network, PSFHSP-Net, based on a modified version of ResNet-18. This network comprises a single convolutional layer and three residual blocks designed to mitigate noise interference and bolster feature extraction capabilities. The model's adaptability was further refined by expanding the shared feature layer into task-specific layers. We assessed its performance against both traditional heavyweight and other lightweight models by evaluating metrics such as F1-score, accuracy (ACC), recall, precision, area under the ROC curve (AUC), model parameter count, and frames per second (FPS). The PSFHSP-Net recorded an ACC of 0.8995, an F1-score of 0.9075, a recall of 0.9191, and a precision of 0.9022. This model surpassed other heavyweight and lightweight models in these metrics. Notably, it featured the smallest model size (1.48 MB) and the highest processing speed (65.7909 FPS), meeting the real-time processing criterion of over 24 images per second. While the AUC of our model was 0.930, slightly lower than that of ResNet34 (0.935), it showed a marked improvement over ResNet-18 in testing, with increases in ACC and F1-score of 0.0435 and 0.0306, respectively. However, precision saw a slight decrease from 0.9184 to 0.9022, a reduction of 0.0162. Despite these trade-offs, the compression of the model significantly reduced its size from 42.64 to 1.48 MB and increased its inference speed by 4.4753 to 65.7909 FPS. The results confirm that the PSFHSP-Net is capable of swiftly and effectively identifying the PSFHSP, thereby facilitating accurate measurements of the angle of progression. This development represents a significant advancement in automating fetal imaging analysis, promising enhanced consistency and reduced operator dependency in clinical settings. Keywords Intrapartum ultrasound · Lightweight network · Ultrasound standard plane detection · Classification · Angle of progression Abbreviations US Ultrasound AoP Angle of progression CNN Convolutional neural network PSFHSP Pubic symphysis-fetal head standard plane PS Pubic symphysis PSFHSP-Net Pubic symphysis-fetal head standard plane-network BN Batch normalization ACC Accuracy ROC Receiver operating characteristic AUC The area under the receiver operating characteristic curve
... ResNet152V2 with GRU model served as the best architecture. The study in [64] presented a deep transfer learning technique using efficient fine-tuning methods for classifying COVID-19 from chest X-ray images. Using two databases, the models achieved average accuracy rates of 95.62%, 96.10%, and 97.61%, with a third model reducing fine-tuning operations. ...
Globally, COVID-19 has impacted severely the healthcare systems and the patients as well. Reverse Transcription-Polymerase Chain Reaction (RT-PCR) tests can be effectively supplemented with computed tomography images. Recent research on CT-based screening found that COVID-19 infection is linked to abnormalities in chest Computed Tomography (CT). However, it is difficult to distinguish these from the general abnormalities that are caused in the lungs. Although COVID-19 RT-PCR testing is exceedingly precise, its sensitivity varies based on the sampling technique and the period as well. Some studies have even shown that RT-PCR testing displays very low COVID-19 sensitivity. This motivated the authors to propose a new deep-learning model called PixNet that can detect positive and negative cases accurately. We compared the effectiveness of the proposed model against several state-of-the-art models trained on CT images. On analysis, it is found that the proposed model displays 96% classification accuracy in diagnosing COVID-19 infection. The proposed algorithm automatically detects the infection owing to COVID-19 from CT scan images, which may be an effective screening tool for Clinicians.
... In the field of medicine, convolutional neural networks (CNNs) are currently the most popular approach for deep learning [18][19][20]. However, CNN models come with certain constraints, one of which is the max-pooling operation. ...
Due to its high infectivity, COVID-19 has rapidly spread worldwide, emerging as one of the most severe and urgent diseases faced by the global community in recent years. Currently, deep learning-based diagnostic methods can automatically detect COVID-19 cases from chest X-ray images. However, these methods often rely on large-scale labeled datasets. To address this limitation, we propose a novel neural network model called CN2A-CapsNet, aiming to enhance the automatic diagnosis of COVID-19 in chest X-ray images through efficient feature extraction techniques. Specifically, we combine CNN with an attention mechanism to form the CN2A model, which efficiently mines relevant information from chest X-ray images. Additionally, we incorporate capsule networks to leverage their ability to understand spatial information, ultimately achieving efficient feature extraction. Through validation on a publicly available chest X-ray image dataset, our model achieved a 98.54% accuracy and a 99.01% recall rate in the binary classification task (COVID-19/Normal) on a six-fold cross-validation dataset. In the three-class classification task (COVID-19/Pneumonia/Normal), it attained a 96.71% accuracy and a 98.34% recall rate. Compared to the previous state-of-the-art models, CN2A-CapsNet exhibits notable advantages in diagnosing COVID-19 cases, specifically achieving a high recall rate even with small-scale datasets.
... Most of these models were trained with the ImageNet dataset [29]. Kaya et al. [25,30] performed experiments on five different pre-trained CNN architectures, viz. VGG16, MobileNetV2, ResNet, InceptionV3, and DenseNet. ...
... MobileNetV2 + Classical fine-tuning, MobileNetV2 + Step fine-tuning, and Mobile-NetV2 + Exponential fine-tuning are the three models developed after using three different fine-tuning mechanisms. The main distinguishing factor in fine-tuning and TL is that in fine-tuning, the entire model is optimised rather than introducing just the classification layer weights for the purpose [30]. ...
This study makes a comprehensive assessment of the predominant Transfer Learning (TL) techniques employed for the classification of COVID-19 cases in Chest X-rays (CXR) images. The methodologies have been selected on the basis of their merits and demerits, suitability, and possible impact on the development of the region being studied. The study examines the various methods of TL employed in the classification of COVID-19 cases with the objective to gain a deeper understanding about all the aspects of these methodologies. It can be of great significance for the researchers and medical professionals in making well-informed decisions about the implementation of these techniques to improve the precision and effectiveness of COVID-19 diagnosis. The practical consequences of these techniques help in early identification of such cases for having a suitable intervention. As many as 48 studies conducted during the period 2020–2023 have been included in the current research work for having an assessment about the problem under investigation. The study has specifically focused on transfer learning-based models utilized for the identification of COVID-19 through CXR pictures. It highlights the challenges posed by dataset dynamics, methodological variations, and performance metrics of different models.
... This work is executed using Raspberry Pi integrated with a workstation and it achieved s detection accuracy of 99.28% [15]. Other related works existing in the literature for the examination of the lung CT slices helped to provide a detection accuracy up to 100% for the chosen image database [16,17]. ...
... Computation of the ensemble probability score is presented in Eqn. (17); ...
Several deep-learning assisted disease assessment schemes (DAS) have been proposed to enhance accurate detection of COVID-19, a critical medical emergency, through the analysis of clinical data. Lung imaging, particularly from CT scans, plays a pivotal role in identifying and assessing the severity of COVID-19 infections. Existing automated methods leveraging deep learning contribute significantly to reducing the diagnostic burden associated with this process. This research aims in developing a simple DAS for COVID-19 detection using the pre-trained lightweight deep learning methods (LDMs) applied to lung CT slices. The use of LDMs contributes to a less complex yet highly accurate detection system. The key stages of the developed DAS include image collection and initial processing using Shannon's thresholding, deep-feature mining supported by LDMs, feature optimization utilizing the Brownian Butterfly Algorithm (BBA), and binary classification through three-fold cross-validation. The performance evaluation of the proposed scheme involves assessing individual, fused, and ensemble features. The investigation reveals that the developed DAS achieves a detection accuracy of 93.80% with individual features, 96% accuracy with fused features, and an impressive 99.10% accuracy with ensemble features. These outcomes affirm the effectiveness of the proposed scheme in significantly enhancing COVID-19 detection accuracy in the chosen lung CT database.
... During the COVID-19 pandemic, many works have been done in areas such as X-rays and CT scan image analysis to detect the disease [12,13,35,36]. The availability of advanced medical equipment and devices is often limited in underdeveloped and developing countries affected by the pandemic, posing challenges in disease diagnosis. ...
... The availability of advanced medical equipment and devices is often limited in underdeveloped and developing countries affected by the pandemic, posing challenges in disease diagnosis. Therefore, it is crucial to develop early and easily accessible detection methods to reduce the mortality rates associated with late detection [13]. The motivation for this study is the need to develop such systems for detecting COVID-19 from cough sounds. ...
... As we navigate the intricate realm of disease classification through sound signals, it is essential to first embark on examining the existing body of research and studies that have paved the way for our exploration. In the field of CAD, several methods have been proposed to detect heart failure [17,38,39], asthma, and COVID-19 using different medical images [12,13] and biomedical signals [23,40] as inputs. These approaches leverage the power of technology to aid in identifying and classifying these disorders. ...
Sound classification has obtained considerable attention in recent years due to its wide range of applications in various fields, such as speech recognition, sound surveillance, music analysis, and environmental monitoring. Because of its success, audio classification can also be employed in medical applications. Coughing is the most common disease symptom, and cough sounds might be used to diagnose them. This research focuses on identifying observable features of cough and classifying them into positive, negative, or symptomatic categories. A novel ensemble learning model based on the super learner (SL) is proposed to diagnose the disease using cough sounds utilizing various audio features such as Frequency Distribution, Time Domain Features, Spectral Features, and Time-Frequency Features. The SL method is a cross-validated approach to stacked generalization, and it can select an optimal learner from a set of learners and improve performance by selecting and merging models using cross-validation. The proposed SL model comprises DT, RF, LR, SVM, ET, and k-NN algorithms. We use the public Coughvid dataset, and the proposed model achieves a correct classification rate for symptomatic cases, which was 90.90%, and the positive predictive value for COVID-19 cases was 84.50%. The SL3 model attains 72%, 78%, 73%, 74.4%, and 78.85% precision, recall, f1-score, accuracy, and average AUC values, respectively. The numerical results show that the proposed model might be implemented to diagnose various other diseases that can be determined from respiratory sounds.
... The new model is named CovidXrayNet which achieved a high classification accuracy of 95.82%. Kaya et al. [8] proposed using MobileNet model with new fine-tuning technique to predict COVID-19 images. They applied their approach on a large dataset with three classes and achieved high accuracy. ...
