Convolutional capsnet: A novel artificial neural network approach to detect COVID-19 disease from X-ray images using capsule networks.

Coronavirus is an epidemic that spreads very quickly. For this reason, it has very devastating effects in many areas worldwide. It is vital to detect COVID-19 diseases as quickly as possible to restrain the spread of the disease. The similarity of COVID-19 disease with other lung infections makes the diagnosis difficult. In addition, the high spreading rate of COVID-19 increased the need for a fast system for the diagnosis of cases. For this purpose, interest in various computer-aided (such as CNN, DNN, etc.) deep learning models has been increased. In these models, mostly radiology images are applied to determine the positive cases. Recent studies show that, radiological images contain important information in the detection of coronavirus. In this study, a novel artificial neural network, Convolutional CapsNet for the detection of COVID-19 disease is proposed by using chest X-ray images with capsule networks. The proposed approach is designed to provide fast and accurate diagnostics for COVID-19 diseases with binary classification (COVID-19, and No-Findings), and multi-class classification (COVID-19, and No-Findings, and Pneumonia). The proposed method achieved an accuracy of 97.24%, and 84.22% for binary class, and multi-class, respectively. It is thought that the proposed method may help physicians to diagnose COVID-19 disease and increase the diagnostic performance. In addition, we believe that the proposed method may be an alternative method to diagnose COVID-19 by providing fast screening.

Raman Spectroscopy of Blood Serum for Essential Thrombocythemia Diagnosis: Correlation with Genetic Mutations and Optimization of Laser Wavelengths.

Essential thrombocythemia (ET) is a type of myeloproliferative neoplasm that increases the risk of thrombosis. To diagnose this disease, the analysis of mutations in the Janus Kinase 2 (JAK2), thrombopoietin receptor (MPL), or calreticulin (CALR) gene is recommended. Disease poses diagnostic challenges due to overlapping mutations with other neoplasms and the presence of triple-negative cases. This study explores the potential of Raman spectroscopy combined with machine learning for ET diagnosis. We assessed two laser wavelengths (785, 1064 nm) to differentiate between ET patients and healthy controls. The PCR results indicate that approximately 50% of patients in our group have a mutation in the JAK2 gene, while only 5% of patients harbor a mutation in the ASXL1 gene. Additionally, only one patient had a mutation in the IDH1 and one had a mutation in IDH2 gene. Consequently, patients having no mutations were also observed in our group, making diagnosis challenging. Raman spectra at 1064 nm showed lower amide, polysaccharide, and lipid vibrations in ET patients, while 785 nm spectra indicated significant decreases in amide II and C-H lipid vibrations. Principal Component Analysis (PCA) confirmed that both wavelengths could distinguish ET from healthy subjects. Support Vector Machine (SVM) analysis revealed that the 800-1800 cm-1 range provided the highest diagnostic accuracy, with 89% for 785 nm and 72% for 1064 nm. These findings suggest that FT-Raman spectroscopy, paired with multivariate and machine learning analyses, offers a promising method for diagnosing ET with high accuracy by detecting specific molecular changes in serum. Principal Component Analysis (PCA) confirmed that both wavelengths could distinguish ET from healthy subjects. Support Vector Machine (SVM) analysis revealed that the 800-1800 cm-1 range provided the highest diagnostic accuracy, with 89% for 785 nm and 72% for 1064 nm. These findings suggest that FT-Raman spectroscopy, paired with multivariate and machine learning analyses, offers a promising method for diagnosing ET with high accuracy by detecting specific molecular changes in serum.

Deep learning approach for prediction of exergy and emission parameters of commercial high by-pass turbofan engines.

Aviation emissions originated from the fuel burn have been hot topics by engineers and policy-makers due to their harmful effects on the environment and thereby human health as well as sustainability. In this study, it is tried that several emission indexes (EIs) involving CO, HC and NOx as well as fuel flow of several commercial aircraft engines (CAEs) are predicted using support vector regression (SVR) and long short-term memory (LSTM) approaches for take-off phase. Moreover, exergo-environmental parameters involving exergy efficiency (ExEFF), wasted exergy ratio (WExR) and environmental effect factor (EEF) pertinent to CAEs are computed employing thermodynamics laws. While establishing the models, rated thrust, by-pass ratio, overall pressure ratio and combustion type of the CAEs are utilized as the model inputs. According to the findings of emission modelling, the coefficient of determination (R2) of EI NOx and EI CO of the CAEs is found as 0.929074 and 0.960277 with SVR, whereas their R2 values are elevated to 0.954878 and 0.989283 with LSTM approach, respectively. However, R2 of EI HC is determined lower with 0.632280 (by SVR) and 0.651749 (by LSTM). On the other hand, exergo-environmental parameters for the CAEs are estimated with high correctness at both models. Namely, R2 of ExEFF and EEF regarding the CAEs are computed as 0.991748 and 0.989067 by SVR; however, these are calculated as 0.994785 and 0.992797 by LSTM method. To model these parameters with low error by using significant design variables as model inputs could help in predicting emission and environmental metrics for new engine designs.

Investigation of the discrimination and characterization of blood serum structure in patients with opioid use disorder using IR spectroscopy and PCA-LDA analysis.

Harmful illicit drug use, such as opioid use disorder (OUD), causes multiple diseases that result in physiological, pathological, and structural changes in serum biochemical parameters based on the period of use. Fourier-transform infrared (FTIR) spectrometry is a noninvasive optical technique that can provide accurate evidence about the biochemical compounds of analytical samples. This technique is based on the detection of functional groups and the spectral analysis of the region of the selected bands, which provides a reliable and accurate tool for evaluating changes in the biochemical parameters of OUD patients. In the present study, the Attenuated Total Reflection (ATR)-FTIR technique and clinical laboratory biochemical results were used to investigate the phospholipid-protein balance in the blood serum of participants with OUD by comparing their data to that of healthy controls. To compare the biochemical laboratory results with serum vibrational spectroscopy, we used infrared (IR) spectroscopy to distinguish the serum of the OUD patients, who had an average duration of use of 7.31 ± 3.8 years (ranging from 6 to 15 years). We aimed to compare the clinical reports with findings from IR spectroscopy coupled with chemometrics analysis, principal component analysis (PCA), and linear discriminant analysis (LDA). The serum samples of the OUD male patients (n = 20) and healthy male individuals (n = 14) were evaluated using FTIR spectroscopy (range of 4000 cm-1 - 400 cm-1). We focused on the areas where the results showed significant band differences and significant chemometric differences at the fingerprint region (1800 cm-1- 900 cm-1), Amide I (1700 cm-1-1600 cm-1), C-H stretching band (3000 cm-1-2800 cm-1), triglyceride (Tg) levels and cholesterol esters (1800 cm-1-1700 cm-1), and total protein region (1700 cm-1 -1350 cm-1). The intensity of these band areas was significantly different (p < 0.01) between OUD patients and healthy controls. Moreover, different bands on the serum spectrum of the OUD patients were explored. The results successfully specified the distinctions between OUD and the healthy controls (HCs). We compared the results with biochemical markers, such as albumin (Alb), Tg, and total cholesterol (Tc) levels of the patients, as well as the data of the healthy subjects obtained from the hospital. Additionally, we found that the Tg, Tc, and Alb levels decreased as the duration of heroin use increased based on the biochemical markers of the OUD patients. The laboratory biochemical reports and the vibrational spectroscopic analysis were correlated. The confidence of specificity, sensitivity, and accuracy was 100%, 92.85%, and 97.06% in the second derivative, respectively. Thus, we demonstrated that IR spectroscopy, multivariate data analysis, and clinical reports are consistent and correlated. Furthermore, FTIR is a simple and readily available diagnostic test that can successfully differentiate the serum samples of OUD patients from those of healthy subjects.

Biochemical assay and spectroscopic analysis of oxidative/antioxidative parameters in the blood and serum of substance use disorders patients. A methodological comparison study.

Substance abuse such as opioids, cannabis, and alcohol causes activation on the immune system and the release of reactive oxygen species (ROS) into the blood and serum. These substances cause an effect on oxidant and antioxidant status in patients with substance abuse. Mainly, wide-open to the ROS are lipids and proteins included blood, which suffers peroxidation. In this study, for the first-time differentiation of the effects of cannabis, alcohol and other synthetic substances on blood and serum samples, were performed. For this purpose, the level of the malondialdehyde (MDA) and glutathione (GSH) in serum and red blood cells, was measured using biochemical assay methods and Fourier Transform InfraRed spectroscopy (FTIR). The results showed, that peroxidation which is dignified as the production of MDA was increased for substance use disorder (SUD) patients (18.01 ±â€¯2.97) compared to the control group (10.75 ±â€¯2.28) (p < 0.001) and for antioxidant capacity, GSH level were significantly increased for SUD patients (p < 0.001). For the discrimination of protein and lipid region obtained from FTIR spectroscopy, we extracted features by principal component analyze (PCA) of protein (1800 cm-1 to 900 cm-1) and lipid (3200 cm-1 to 2800 cm-1) regions for blood and serum samples collected from patients with different types of SUD and healthy control (HC) participants. For the consideration of lipid oxidation, lipid saturation, lipid desaturation and protein aggregation the peak heights at 1740 cm-1 to 2960 cm-1, 2920 cm-1 to 2960 cm-1, 3012 cm-1 to 2960 cm-1, and 1630 cm-1 to 1650 cm-1 regions were studied for SUD and HC. Moreover, more visible changes were noticed for proteins region, than for lipids. The most notice structural changes were observed in amide II in serum spectra. Then we classified protein and lipid region's PCA results of blood and serum by Linear discriminant analysis (LDA) and Support vector machine (SVM). Confidence of the specificity, sensitivity and accuracy of blood and serum were obtained as 100%, 100% and 100% individually. This is the first comparative study on the spectrochemical tool and biochemical assay on SUD. Our results presented 100% discrimination of disorder region compared to healthy subjects.

Is it possible to detect cerebral dominance via EEG signals by using deep learning?

Each brain hemisphere is dominant for certain functions such as speech. The determination of speech laterality prior to surgery is of paramount importance for accurate risk prediction. In this study, we aimed to determine speech laterality via EEG signals by using noninvasive machine learning techniques. The retrospective study included 67 subjects aged 18-65 years who had no chronic diseases and were diagnosed as healthy based on EEG examination. The subjects comprised 35 right-hand dominant (speech center located in the left hemisphere) and 32 left-hand dominant individuals (speech center located in the right hemisphere). A spectrogram was created for each of the 18 EEG channels by using various Convolutional Neural Networks (CNN) architectures including VGG16, VGG19, ResNet, MobileNet, NasNet, and DenseNet. These architectures were used to extract features from the spectrograms. The extracted features were classified using Support Vector Machines (SVM) and the classification performances of the CNN models were evaluated using Area Under the Curve (AUC). Of all the CNN models used in the study, VGG16 had a higher AUC value (0.83 ±â€¯0.05) in the determination of speech laterality compared to all other models. The present study is a pioneer investigation into the determination of speech laterality via EEG signals with machine learning techniques, which, to our knowledge, has never been reported in the literature. Moreover, the classification results obtained in the study are promising and lead the way for subsequent studies though not practically feasible.

A deep learning-based decision support system for diagnosis of OSAS using PTT signals.

Sleep disorders, which negatively affect an individual's daily quality of life, are a common problem for most of society. The most dangerous sleep disorder is obstructive sleep apnea syndrome (OSAS), which manifests itself during sleep and can cause the sudden death of patients. Many important parameters related to the diagnosis and treatments of such sleep disorders are simultaneously examined. This process is exhausting and time-consuming for experts and also requires experience; thus, it can cause difference of opinion among experts. Because of this, automatic sleep staging systems have been designed. In this study, a decision support system was developed to determine OSAS patients. In the developed decision support system, unlike in the available published literature, patient and healthy individual classification was performed using only the Pulse Transition Time (PTT) parameter rather than other parameters obtained from polysomnographic data like ECG (Electrocardiogram), EEG (Electroencephalography), carbon dioxide measurement and EMG (Electromyography). The suggested method can perform feature extraction from PTT signals by means of a deep-learning method. AlexNet and VGG-16, which are two Convolutional Neural Network (CNN) models, have been used for feature extraction. With the features obtained, patients and healthy individuals were classified by the Support Vector Machine (SVM) and the k-nearest neighbors (k-NN) algorithms. When the performance of the study was compared with other studies in published literature, it was seen that satisfactory results were obtained.