Secure Convolutional Neural Network-based Internet-of-Healthcare Applications

Convolutional neural networks (CNNs) have gained popularity for Internet-of-Healthcare (IoH) applications such as medical diagnostics. However, new research shows that adversarial attacks with slight imperceptible changes can undermine deep neural network techniques in healthcare. This raises questions regarding the safety of deploying these IoH devices in clinical situations. In this paper, we review the techniques used in fighting against cyber-attacks. Then, we propose to study the robustness of some well-known CNN architectures’belonging to sequential, parallel, and residual families, such as LeNet5, MobileNetV1, VGG16, ResNet50, and InceptionV3 against fast gradient sign method (FGSM) and projected gradient descent (PGD) attacks, in the context of classification of chest radiographs (X-rays) based on the IoH application. Finally, we propose to improve the security of these CNN structures by studying standard and adversarial training. The results show that, among these models, smaller models with lower computational complexity are more secure against hostile threats than larger models that are frequently used in IoH applications. In contrast, we reveal that when these networks are learned adversarially, they can outperform standard trained networks. The experimental results demonstrate that the model performance breakpoint is represented by γ= 0.3 with a maximum loss of accuracy tolerated at 2%. Author

CT-Scans Images Segmentation for COVID-19 Detection Based CNN Models

Covid-19 disease has been known as a spreaded epidemic across the whole world that affects millions of people, causing deaths and catastrophic effects. For this reason, Computer Aided Diagnosis System (CAD), consists to be a crucial step using deep learning algorithms. In this context, a CNN network has been proposed using two optimizers networks such as Adam and Adadelta. The whole system is implemented on GPU with the aim to speed up the implementation time process. Then to have a medical real application which automatically detect the covid-19 class from CT scans images. Segmentations results achieved in terms of training and validation accuracies are 99.54% 99.65% respectively, outperforming the state of the art. Moreover, the predicted segmented images shows excellent results in terms of Mean Square Error reaching to 0.009, which is close to zero, compared to the ground truth. As a result, a medical real time application is achieved for Covid-19 class segmentation in a short time process. © 2021 IEEE.

Real-Time Application for Covid-19 Class Detection based CNN Architecture

Covid-19 disease has been known as a spreaded epidemic across the whole world that affects millions of people, causing deaths and catastrophic effects. For this reason, Computer Aided Diagnosis System (CAD), consists to be a crucial step using deep learning algorithms. In this context, a CNN network has been proposed using two optimizers networks such as Rmsprop and SGD with momentum.the whole system is implemented on both CPU and GPU with the aim to speed up the implementation time process. Then to have a medical real application which automatically detect the covid-19 class from X-rays images of chest. Classification results achieved in terms of accuracy, specificity and sensitivity 99.22%, 99.65% and 99.45% respectively, outperforming the state of the art. As a result, a medical real time application is achieved for Covid-19 class detection in a short time process. © 2021 IEEE.

COVID-19 Recognition Based on Patient's Coughing and Breathing Patterns Analysis: Deep Learning Approach

The World Health Organization has declared that the new Coronavirus disease (Covid-19) has become a pandemic since March 2020. It consists of an emerging viral infection with respiratory swelling that can progress to atypical pneumonia. In fact, experts stress the early detection importance of those infected with COVID-19 virus. In this way, the infected patients will be isolated from others, and then prevent the virus spread. However, prompt assessment of breathing patterns is important for many medical emergencies. We present, in this paper, a deep learning technique-based COVID-19 cough and breath analysis that can recognize positive COVID-19 cases from both negative and healthy COVID-19 cough and breath recorded on smartphones or wearable sensors. Firstly, audio signals, as well as cough and breath, will be preprocessed to remove noise. After that, deep features will be extracted using the deep Long Term Short Memory (LSTM) model. Finally, the recognition step will be performed exploiting extracted audio features. Numerical results prove the efficiency of the proposed deep model in terms of high accuracy level and low loss value compared to the other techniques. © 2021 FRUCT.