Automatic Sleep Staging Using BiRNN with Data Augmentation and Label Redirection

Sleep staging has always been a hot topic in the field of sleep medicine, and it is the cornerstone of research on sleep problems. At present, sleep staging heavily relies on manual interpretation, which is a time-consuming and laborious task with subjective interpretation factors. In this paper, we propose an automatic sleep stage classification model based on the Bidirectional Recurrent Neural Network (BiRNN) with data bundling augmentation and label redirection for accurate sleep staging. Through extensive analysis, we discovered that the incorrect classification labels are primarily concentrated in the transition and nonrapid eye movement stage I (N1). Therefore, our model utilizes a sliding window input to enhance data bundling and an attention mechanism to improve feature enhancement after label redirection. This approach focuses on mining latent features during the N1 and transition periods, which can further improve the network model's classification performance. We evaluated on multiple public datasets and achieved an overall accuracy rate of 87.3%, with the highest accuracy rate reaching 93.5%. Additionally, the network model's macro F1 score reached 82.5%. Finally, we used the optimal network model to study the impact of different EEG channels on the accuracy of each sleep stage.

A multivariate spatiotemporal model for tracking COVID-19 incidence and death rates in socially vulnerable populations.

Recent studies have produced inconsistent findings regarding the association between community social vulnerability and COVID-19 incidence and death rates. This inconsistency may be due, in part, to the fact that these studies modeled cases and deaths separately, ignoring their inherent association and thus yielding imprecise estimates. To improve inferences, we develop a Bayesian multivariate negative binomial model for exploring joint spatial and temporal trends in COVID-19 infections and deaths. The model introduces smooth functions that capture long-term temporal trends, while maintaining enough flexibility to detect local outbreaks in areas with vulnerable populations. Using multivariate autoregressive priors, we jointly model COVID-19 cases and deaths over time, taking advantage of convenient conditional representations to improve posterior computation. As such, the proposed model provides a general framework for multivariate spatiotemporal modeling of counts and rates. We adopt a fully Bayesian approach and develop an efficient posterior Markov chain Monte Carlo algorithm that relies on easily sampled Gibbs steps. We use the model to examine incidence and death rates among counties with high and low social vulnerability in the state of Georgia, USA, from 15 March to 15 December 2020.

Boosting Masked Face Recognition with Multi-Task ArcFace

In this article, we tackle the recognition of faces wearing surgical masks. Surgical masks have become a necessary piece of daily apparel because of the COVID-19-related worldwide health problem. Modern face recognition models are in trouble because they were not made to function with masked faces. Furthermore, in order to stop the infection from spreading, apps capable of detecting if the individuals are wearing masks are also required. To address these issues, we present an end-to-end approach for training face recognition models based on the ArcFace architecture, including various changes to the backbone and loss computation. We also use data augmentation to generate a masked version of the original dataset and mix them on the fly while training. Without incurring any additional computational costs, we modify the chosen network to output also the likelihood of wearing a mask. Thus, the face recognition loss and the mask-usage loss are merged to create a new function known as Multi-Task ArcFace (MTArcFace). The conducted experiments demonstrate that our method outperforms the baseline model results when faces with masks are considered, while achieving similar metrics on the original dataset. In addition, it obtains a 99.78% of mean accuracy in mask-usage classification. © 2022 IEEE.

EXPLORING THE IMPACT OF IMAGE ENHANCEMENT AND DATA AUGMENTATION TECHNIQUES ON LUNG DETECTION IN CHEST RADIOGRAPHY IMAGES

The health and wellbeing of people all over the world are being severely impacted by the ongoing COVID-19 pandemic. One of the most important ways to check for COVID-19 is chest radiography, so ensuring that infected people undergo this test is crucial. This research set out to assess the efficacy of various image enhancement and data augmentation techniques for use with digital chest X-Rays in the detection of COVID-19 patients. White-balance correction (WB) and contrast-limited adaptive histogram equalization (CLAHE) were the two methods used to improve the images. These two technologies have also been applied to examine this impact on COVID-19 discrimination. Also, Data was augmented in two distinct ways, using a different set of techniques and combining it with image enhancement techniques. Transfer learning was used to compare image classification models pre-trained on the ImageNet dataset to well-known deep learning architectures. Our models were evaluated and compared using the novel-combined chest X-Ray datasets. We observed that the VGG-16 model outperforms other models with an accuracy of 98% when image WB and CLAHE are used together. Due to their superior performance, these pre-trained models can greatly improve the speed and accuracy of COVID-19 diagnosis. © 2023, International Organization on 'Technical and Physical Problems of Engineering'. All rights reserved.

Adversarial Approaches to Tackle Imbalanced Data in Machine Learning

Real-world applications often involve imbalanced datasets, which have different distributions of examples across various classes. When building a system that requires a high accuracy, the performance of the classifiers is crucial. However, imbalanced datasets can lead to a poor classification performance and conventional techniques, such as synthetic minority oversampling technique. As a result, this study proposed a balance between the datasets using adversarial learning methods such as generative adversarial networks. The model evaluated the effect of data augmentation on both the balanced and imbalanced datasets. The study evaluated the classification performance on three different datasets and applied data augmentation techniques to generate the synthetic data for the minority class. Before the augmentation, a decision tree was applied to identify the classification accuracy of all three datasets. The obtained classification accuracies were 79.9%, 94.1%, and 72.6%. A decision tree was used to evaluate the performance of the data augmentation, and the results showed that the proposed model achieved an accuracy of 82.7%, 95.7%, and 76% on a highly imbalanced dataset. This study demonstrates the potential of using data augmentation to improve the classification performance in imbalanced datasets.

A linear noise approximation for stochastic epidemic models fit to partially observed incidence counts.

Stochastic epidemic models (SEMs) fit to incidence data are critical to elucidating outbreak dynamics, shaping response strategies, and preparing for future epidemics. SEMs typically represent counts of individuals in discrete infection states using Markov jump processes (MJPs), but are computationally challenging as imperfect surveillance, lack of subject-level information, and temporal coarseness of the data obscure the true epidemic. Analytic integration over the latent epidemic process is impossible, and integration via Markov chain Monte Carlo (MCMC) is cumbersome due to the dimensionality and discreteness of the latent state space. Simulation-based computational approaches can address the intractability of the MJP likelihood, but are numerically fragile and prohibitively expensive for complex models. A linear noise approximation (LNA) that approximates the MJP transition density with a Gaussian density has been explored for analyzing prevalence data in large-population settings, but requires modification for analyzing incidence counts without assuming that the data are normally distributed. We demonstrate how to reparameterize SEMs to appropriately analyze incidence data, and fold the LNA into a data augmentation MCMC framework that outperforms deterministic methods, statistically, and simulation-based methods, computationally. Our framework is computationally robust when the model dynamics are complex and applies to a broad class of SEMs. We evaluate our method in simulations that reflect Ebola, influenza, and SARS-CoV-2 dynamics, and apply our method to national surveillance counts from the 2013-2015 West Africa Ebola outbreak.

Biometric Russian Audio-Visual Extended MASKS (BRAVE-MASKS) Corpus: Multimodal Mask Type Recognition Task

In this paper, we present a new multimodal corpus called Biometric Russian Audio-Visual Extended MASKS (BRAVE-MASKS), which is designed to analyze voice and facial characteristics of persons wearing various masks, as well as to develop automatic systems for bimodal verification and identification of speakers. In particular, we tackle the multimodal mask type recognition task (6 classes). As a result, audio, visual and multimodal systems were developed, which showed UAR of 54.83%, 72.02% and 82.01%, respectively, on the Test set. These performances are the baseline for the BRAVE-MASKS corpus to compare the follow-up approaches with the proposed systems.

Convolutional Neural Networks for Detection of COVID-19 from Chest X-Rays

The coronavirus (COVID-19) pandemic was rapid in its outbreak, and the contagion of the virus led to an extensive loss of life globally. This study aims to propose an efficient and reliable means to differentiate between chest x-rays indicating COVID-19 and other lung conditions. The proposed methodology involved combining deep learning techniques such as data augmentation, CLAHE image normalization, and transfer learning with eight pre-trained networks. The highest performing networks for binary, 3-class (normal vs. COVID-19 vs. viral pneumonia) and 4-class classifications (normal vs. COVID-19 vs. lung opacity vs. viral pneumonia) were MobileNetV2, InceptionResNetV2, and MobileNetV2, achieving accuracies of 97.5%, 96.69%, and 92.39%, respectively. These results outperformed many state-of-the-art methods conducted to address the challenges relating to the detection of COVID-19 from chest x-rays. The method proposed can serve as a basis for a computer-aided diagnosis (CAD) system to ensure that patients receive timely and necessary care for their respective illnesses. Copyright © 2022, IGI Global.

Automated COVID-19 Detection from WBC-DIFF Scattergram Images with Hybrid CNN Model Using Feature Selection

In the medical diagnosis such as WBC (white blood cell), the scattergram images show the relationships between neutrophils, eosinophils, basophils, lymphocytes, and monocytes cells in the blood. For COVID-19 detection, the distributions of these cells differ in healthy and COVID-19 patients. This study proposes a hybrid CNN model for COVID-19 detection using scatter images obtained from WBC sub (differential-DIFF) parameters instead of CT or X-Ray scans. As a data set, the scattergram images of 335 COVID-19 suspects without chronic disease, collected from the biochemistry department of Elazig Fethi Sekin City Hospital, are examined. At first, the data augmentation is performed by applying HSV(Hue, Saturation, Value) and CIE-1931(Commission Internationale de l'éclairage) conversions. Thus, three different image large sets are obtained as a result of raw, CIE-1931, and HSV conversions. Secondly, feature extraction is applied by giving these images as separate inputs to the CNN model. Finally, the ReliefF feature extraction algorithm is applied to determine the most dominant features in feature vectors and to determine the features that maximize classification accuracy. The obtaining feature vector is classified with high-performance SVM in binary classification. The overall accuracy is 95.2%, and the F1-Score is 94.1%. The results show that the method can successfully detect COVID -19 disease using scattergram images and is an alternative to CT and X-Ray scans.

The Optimization of COVID-19 Detection Based on DesNet Model

In recent years, with the pandemic of COVID-19, how to identify the positive cases of COVID-19 accurately and rapidly from patients has become the key to block the spread of the epidemic and assist clinical diagnosis. In this paper, a COVID-19 detection model was constructed for the purpose to identify the positive cases from patients with other lung diseases as well as the normal using the chest X-ray images. The basic structure of the detection system is a CNN model based on DesNet with some optimization algorithms and the accuracy has reached 94.2%. We also applied three multi-sample data augmentation methods: SMOTE, mixup and CutMix to the model to analyze their performance. By applying these methods, the model finally reached 97.9% on test set and showed a good generalization on other datasets, which could reach over 80% without extra training. The results show that using transfer learning and some muli-sample data augmentation methods can significantly improve the accuracy and overcome overfitting problem of fewshot learning, while others may not be so effective. © 2023 IEEE.

Identification and Detecting COVID-19 from X-Ray Images Using Convolutional Neural Networks

The sudden increase of COVID-19 patients is alarming, and it requires quick diagnosis in a quick time. PCR testing is one of the most used methods to test and diagnose COVID, which is time-consuming. In this paper, we present an end-to-end technique that can detect COVID-19 using chest X-ray scans. We have trained and optimized a convolutional neural network (ConvNet), which was trained on a large COVID-19 dataset. We have performed a series of experiments on a number of different architectures. We have chosen the best performing network architecture and then carried on a series of additional experiments to find the optimal set of hyper-parameters and show and justify a number of data augmentation strategies that have allowed us to enhance our performance on the test set greatly. Our final trained ConvNet has managed to obtain a test accuracy of 97.89%. This high accuracy and very fast test speed can be beneficial to get quick COVID test results for further treatment. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

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.

Prediction Model for Inclusive Finance Development Considering the Impact of COVID-19: The Case of China

Models in previous studies about inclusive finance often include economic data while excludes public online statements. In this paper Random Forest Regression (RFR) model is trained on the annual influencing factors and annual financial inclusion index to predict quarterly financial inclusion index by the quarterly influencing factors to expand the size of data. Then, BOW model tf-idf algorithm is used to convert COVID-19 – loan related online statements into word vectors. Lastly, these influencing factors of different lag periods are passed into the RFR model to compare their performance. Result of models shows that there is impact the epidemic has on the development of inclusive finance, and the lag period of the impact opinion texts on financial inclusion is 2 quarters. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

TN-GAN-Based Pet Behavior Prediction through Multiple-Dimension Time-Series Augmentation.

Behavioral prediction modeling applies statistical techniques for classifying, recognizing, and predicting behavior using various data. However, performance deterioration and data bias problems occur in behavioral prediction. This study proposed that researchers conduct behavioral prediction using text-to-numeric generative adversarial network (TN-GAN)-based multidimensional time-series augmentation to minimize the data bias problem. The prediction model dataset in this study used nine-axis sensor data (accelerometer, gyroscope, and geomagnetic sensors). The ODROID N2+, a wearable pet device, collected and stored data on a web server. The interquartile range removed outliers, and data processing constructed a sequence as an input value for the predictive model. After using the z-score as a normalization method for sensor values, cubic spline interpolation was performed to identify the missing values. The experimental group assessed 10 dogs to identify nine behaviors. The behavioral prediction model used a hybrid convolutional neural network model to extract features and applied long short-term memory techniques to reflect time-series features. The actual and predicted values were evaluated using the performance evaluation index. The results of this study can assist in recognizing and predicting behavior and detecting abnormal behavior, capacities which can be applied to various pet monitoring systems.

AI-based wavelet and stacked deep learning architecture for detecting coronavirus (COVID-19) from chest X-ray images.

A novel coronavirus (COVID-19), belonging to a family of severe acute respiratory syndrome coronavirus 2 (SARs-CoV-2), was identified in Wuhan city, Hubei, China, in November 2019. The disease had already infected more than 681.529665 million people as of March 13, 2023. Hence, early detection and diagnosis of COVID-19 are essential. For this purpose, radiologists use medical images such as X-ray and computed tomography (CT) images for the diagnosis of COVID-19. It is very difficult for researchers to help radiologists to do automatic diagnoses by using traditional image processing methods. Therefore, a novel artificial intelligence (AI)-based deep learning model to detect COVID-19 from chest X-ray images is proposed. The proposed work uses a wavelet and stacked deep learning architecture (ResNet50, VGG19, Xception, and DarkNet19) named WavStaCovNet-19 to detect COVID-19 from chest X-ray images automatically. The proposed work has been tested on two publicly available datasets and achieved an accuracy of 94.24% and 96.10% on 4 classes and 3 classes, respectively. From the experimental results, we believe that the proposed work can surely be useful in the healthcare domain to detect COVID-19 with less time and cost, and with higher accuracy.

Ensemble deep honey architecture for COVID-19 prediction using CT scan and chest X-ray images.

Recently, the infectious disease COVID-19 remains to have a catastrophic effect on the lives of human beings all over the world. To combat this deadliest disease, it is essential to screen the affected people quickly and least inexpensively. Radiological examination is considered the most feasible step toward attaining this objective; however, chest X-ray (CXR) and computed tomography (CT) are the most easily accessible and inexpensive options. This paper proposes a novel ensemble deep learning-based solution to predict the COVID-19-positive patients using CXR and CT images. The main aim of the proposed model is to provide an effective COVID-19 prediction model with a robust diagnosis and increase the prediction performance. Initially, pre-processing, like image resizing and noise removal, is employed using image scaling and median filtering techniques to enhance the input data for further processing. Various data augmentation styles, such as flipping and rotation, are applied to capable the model to learn the variations during training and attain better results on a small dataset. Finally, a new ensemble deep honey architecture (EDHA) model is introduced to effectively classify the COVID-19-positive and -negative cases. EDHA combines three pre-trained architectures like ShuffleNet, SqueezeNet, and DenseNet-201, to detect the class value. Moreover, a new optimization algorithm, the honey badger algorithm (HBA), is adapted in EDHA to determine the best values for the hyper-parameters of the proposed model. The proposed EDHA is implemented in the Python platform and evaluates the performance in terms of accuracy, sensitivity, specificity, precision, f1-score, AUC, and MCC. The proposed model has utilized the publicly available CXR and CT datasets to test the solution's efficiency. As a result, the simulated outcomes showed that the proposed EDHA had achieved better performance than the existing techniques in terms of Accuracy, Sensitivity, Specificity, Precision, F1-Score, MCC, AUC, and Computation time are 99.1%, 99%, 98.6%, 99.6%, 98.9%, 99.2%, 0.98, and 820 s using the CXR dataset.

Deep learning based COVID-19 detection using medical images: Is insufficient data handled well?

The deep learning is a prominent method for automatic detection of COVID-19 disease using medical dataset. This paper aims to give the perspective on the data insufficiency issue that exists in COVID-19 detection associated with deep learning. The extensive study on the available datasets comprising CT and X-ray images are presented in this paper, which can be very much useful in the context of deep learning framework for COVID-19 detection. Moreover, various data handling techniques that are very essential in deep learning models are discussed in detail. Advanced data handling techniques and approaches to modify deep learning models are suggested to handle the data insufficiency problem in deep learning based COVID-19 detection.

Segmentation Algorithm for COVID-19 Infections using Joint Mix and Context Encoder

Automatic image segmentation is critical for medical image segmentation. For example, automatic segmentation of infection area of COVID-19 before and after diagnosis and treatment can help us automatically analyze the diagnosis and treatment effect. The existing algorithms do not solve the problems of insufficient data and insufficient feature extraction at the same time. In this paper, we propose a new data augmentation algorithm to handle the insufficient data problem, named Joint Mix;we utilize an improved U-Net with context encoder to enhance the feature extraction ability. Experiments in the segmentation of COVID-19 infection region using CT images demonstrate its effectiveness. © 2022 IEEE.

DACov: a deeper analysis of data augmentation on the computed tomography segmentation problem

Due to the COVID-19 global pandemic, computer-assisted diagnoses of medical images have gained much attention, and robust methods of semantic segmentation of Computed Tomography (CT) images have become highly desirable. In this work, we present a deeper analysis of how data augmentation techniques improve segmentation performance on this problem. We evaluate (Formula presented.) traditional augmentation techniques on five public datasets. Six different probabilities of applying each augmentation technique on an image were evaluated. We also assess a different training methodology where the training subsets are combined into a single larger set. All networks were evaluated through a (Formula presented.) -fold cross-validation strategy, resulting in over (Formula presented.) experiments. We also propose a novel data augmentation technique based on Generative Adversarial Networks (GANs) to create new healthy and unhealthy lung CT images, evaluating four variations of our approach with the same six probabilities of the traditional methods. Our findings show that GAN-based techniques and spatial-level transformations are the most promising for improving the learning of deep models on this problem, with the StarGAN v2 + F with a probability of (Formula presented.) achieving the highest F-score value on the Ricord1a dataset in the unified training strategy. Our code is publicly available at https://github.com/VRI-UFPR/DACov2022.Copyright © 2023 Informa UK Limited, trading as Taylor & Francis Group.

HateNet: A Graph Convolutional Network Approach to Hate Speech Detection

The COVID-19 pandemic has caused hate speech on online social networks to become a growing issue in recent years, affecting millions. Our work aims to improve automatic hate speech detection to prevent escalation to hate crimes. The first c hallenge i n h ate s peech r esearch i s t hat e xisting datasets suffer from quite severe class imbalances. The second challenge is the sparsity of information in textual data. The third challenge is the difficulty i n b alancing t he t radeoff b etween utilizing semantic similarity and noisy network language. To combat these challenges, we establish a framework for automatic short text data augmentation by using a semi-supervised hybrid of Substitution Based Augmentation and Dynamic Query Expansion (DQE), which we refer to as SubDQE, to extract more data points from a specific c lass f rom T witter. W e a lso p ropose the HateNet model, which has two main components, a Graph Convolutional Network and a Weighted Drop-Edge. First, we propose a Graph Convolutional Network (GCN) classifier, using a graph constructed from the thresholded cosine similarities between tweet embeddings to provide new insights into how ideas are connected. Second, we propose a weighted Drop-Edge based stochastic regularization technique, which removes edges randomly based on weighted probabilities assigned by the semantic similarities between Tweets. Using 3 different SubDQE-augmented datasets, we compare our HateNet model using eight different tweet embedding methods, six other baseline classification models, and seven other baseline data augmentation techniques previously used in the realm of hate speech detection. Our results show that our proposed HateNet model matches or exceeds the performance of the baseline models, as indicated by the accuracy and F1 score. © 2022 IEEE.