A study is conducted to investigate the use of a parametric intensity model for the process of image classification in biomedical microwave tomography (MWT). This process allows for extracting ...structural information about an object-of-interest (OI), which can be incorporated as prior information in an inversion algorithm. The parametric intensity model is based on a supervised Gaussian probabilistic model. The generated intensity model is used to classify three cross-sectional MWT images of human lower leg models. The classification is based on a Bayesian decision classifier. The resulting segments are used to extract structural information about the legs' contour.
Brain tumor segmentation from magnetic resonance (MR) images can have a great impact on improving diagnostics, growth rate prediction, and treatment planning. In this paper, we provide a comparative ...study of four well-known segmentation algorithms, namely k-means clustering, histogram thresholding (Otsu), fuzzy c-means thresholding, and region growing. For the region growing algorithm, the seed selection process is automated and enhanced by preprocessing the images and approximating the tumor regions using initial clustering and/or thresholding approaches. The evaluation and comparison of the algorithms is conducted using a data-set of T1-Weighted Contrast-Enhanced magnetic resonance imaging (MRI) brain images. Ground truth tumor images were provided by three experienced radiologists and are used in the evaluation process. Results showed that the enhanced region growing method had the highest mean dice similarity coefficient with a score of 0.87, and the lowest under-segmentation rate (17.46%). The fuzzy c-means thresholding method had the lowest over-segmentation rate (0.03%). This study serves as a baseline for other advanced tumor segmentation studies such as the ones using the emergent machine learning approaches.
Assessing fetal well-being using conventional tools requires skilled clinicians for interpretation and can be susceptible to noise interference, especially during lengthy recordings or when maternal ...effects contaminate the signals. In this study, we present a novel transformer-based deep learning model called fetal heart sounds U-Net Transformer (FHSU-NETR) for automated extraction of fetal heart activity from raw phonocardiography (PCG) signals. The model was trained using a realistic synthetic dataset and validated on data recorded from 20 healthy mothers at the pregnancy outpatient clinic of Tohoku University Hospital, Japan. The model successfully extracted fetal PCG signals; achieving a heart rate mean difference of −1.5 bpm compared to the ground-truth calculated from fetal electrocardiogram (ECG). By leveraging deep learning, FHSU-NETR would facilitates timely interpretation of lengthy PCG recordings while reducing the heavy reliance on medical experts; thereby enhancing the efficiency in clinical practice.
In this paper, a preliminary numerical study is conducted to investigate the use of microwave tomography in monitoring bone density in human lower limbs. By monitoring bone density, the effectiveness ...of Vitamin D treatment can be evaluated for Osteoporosis patients. In microwave tomography, the leg is radiated with non-ionizing low-power electromagnetic signals with scattered electric fields measured at several locations surrounding the leg. Within the framework of inverse scattering problems, the measured fields are used as inputs for an optimization algorithm to estimate the location and electrical properties inside the human leg. In this work, a two-dimensional cross-sectional model of a human leg is created and simulated using a finite-element method where the transverse magnetic approximation is applied. The synthetic results are then inverted using a finite-element contrast source inversion method. The results show that variations in bone density effect the results of the inversion algorithm.
In this paper, a numerical study is conducted to provide guidelines towards designing a wearable microwave tomography (MWT) system for medical imaging. The main objective is to ease the process of ...imaging human organs as well as improve the overall image reconstruction of the utilized inversion algorithm. In comparison to conventional MWT systems currently available, a wearable system uses structural information about the object-of-interest (OI) with ease in its design. The information is used to relocate the antennas and the inversion algorithm image domain. In addition, a wearable system uses ultrasound gel as a coupling medium, which is an alternative to a liquid matching medium. As an example, the proposed guidelines are applied on the use of MWT for imaging human lower limbs. The initial results of a simulated wearable system show a better estimation of the relative complex permittivity of different tissues within a OI in comparison to a conventional MWT system.
Emotions play a pivotal role in the individual's overall physical health. Therefore, there has been a steadily increasing interest towards emotion recognition in conversation (ERC). In this work, we ...propose bidirectional long short term memory (Bi-LSTM), convolutional neural network (CNN), and CNN-BiLSTM based models to predict the emotional climate established during the conversation by peers. Their speech signals across their conversation are analyzed using Mel frequency cepstral coefficients (MFCCs) that are then fed to the Bi-LSTM, CNN and CNN-BiLSTM models to predict the valence and arousal emotional climate cues. The proposed approach was tested on a publicly available dataset, namely K-EmoCon, that includes emotion labeling and peers' speech signals, during their conversation. The obtained results show that Bi-LSTM, CNN and CNN-BiLSTM models achieved a classification accuracy (arousal/valence) of 67.5%/57.7%, 73.3%/66.9%, and 75.1%/68.3%, respectively. These encouraging results show that a combination of deep learning schemes could increase the classification accuracy and provide efficient emotional climate recognition in naturalistic conversation environments.
Heart failure refers to the inability of the heart to pump enough amount of blood to the body. Nearly 7 million people die every year because of its complications. Current gold-standard screening ...techniques through echocardiography do not incorporate information about the circadian rhythm of the heart and clinical information of patients. In this vein, we propose a novel approach to integrate 24-hour heart rate variability (HRV) features and patient profile information in a single multi-parameter and color-coded polar representation. The proposed approach was validated by training a deep learning model from 7,575 generated images to predict heart failure groups, i.e., preserved, mid-range, and reduced left ventricular ejection fraction. The developed model had overall accuracy, sensitivity, and specificity of 93%, 88%, and 95%, respectively. Moreover, it had a high area under the receiver operating characteristics curve (AUROC) of 0.88 and an area under the precision-recalled curve (AUPR) of 0.79. The novel approach proposed in this study suggests a new protocol for assessing cardiovascular diseases to act as a complementary tool to echocardiography as it provides insights on the circadian rhythm of the heart and can be potentially personalized according to patient clinical profile information.Clinical relevance- Implementing polar representations with deep learning in clinical settings to supplement echocardiography leverages continuous monitoring of the heart's circadian rhythm and personalized cardiovascular medicine while reducing the burden on medical practitioners.
Cardiac auscultation through phonocardiogram (PCG) is still the most commonly used approach for evaluating the mechanical functionality of the heart when diagnosing congenital heart disease. Despite ...of its time- and cost-effectiveness, it is still limited due to the extensive need for clinical expertise for interpretation. In this study, we propose the use of ensemble transformer-based neural networks to aid in the detection of heart murmur in PCG recordings and for the prediction of clinical outcomes of patients as part of George B. Moody PhysioNet 2022 Challenge. Our team, Care4MyHeart, developed an approach that transforms the raw PCG recordings into wavelet power features signals for the use within the proposed deep learning models. We have achieved a maximum accuracy of 0.855, 0.761, and 0.757 for murmur detection in the training, hidden validation, and hidden testing datasets, respectively. In addition, we had an overall clinical outcome cost of 9980, 11490, and 14410 for the three datasets, respectively. Our team was ranked 6th/40 for murmur detection and 29th/40 for clinical outcome predictions. We had the lowest clinical outcome cost on the validation set of 9737 with a murmur detection score of 0.730 when reducing the number of features used to train the models.
Emotion recognition in conversations using artificial intelligence (AI) has recently gained a lot of attention, as it can provide additional emotion cues that can be correlated with human social ...behavior. An extension towards an AI-based emotional climate (EC) recognition, i.e., the recognition of the joint emotional atmosphere dynamically created and perceived by the peers throughout a conversation, is proposed here. In our approach, namely MLBispeC (Machine Learning Based Bispectral Classification), the peers' speech signals during their conversation are subjected to time-windowed bispectral analysis, allowing for feature extraction related to dynamic harmonics nonlinear interactions. In addition, peers' affect dynamics, derived from their same time-windowed emotion labeling, are combined to form an extended feature vector, inputted into two well-known machine learning classifiers (Support Vector Machine, K-Nearest Neighbor). MLBispeC was evaluated on the Interactive Emotional Dyadic Motion Capture (IEMOCAP) open access dataset, which contains 2D emotions, i.e., Arousal (A) and valence (V) that are divided into (low/high) classes. The experimental results have shown that MLBispeC outperforms previous state-of-the-art techniques, achieving an accuracy of 0.826A/0.754V, sensitivity of 0.864A/0.774V, and area under the curve (AUC) of 0.821A/0.799V. This demonstrates the effectiveness of MLBispeC to objectively recognize peers' EC during their conversation, allowing for insights into their emotional and social interactions.Clinical relevance-Unobtrusive, objective and dynamic recognition of the EC built during peers' conversation can scaffold effective assessment of patients with physiological, psychological, and mental diseases, at various age ranges (children, adults, and older adults)
This study explored how 24-hour Heart Rate Variability (HRV) features differentiate amongst Coronary Artery Disease (CAD) patients with "at risk", "borderline", and "normal" Left Ventricular Ejection ...Fraction (LVEF). Hourly segmentation of heart rate signals was completed by Cosinor Analysis fitting. Time, frequency, and non-linear HRV features were estimated for each hour and averaged across all CAD patients for each group. Statistical analysis to identify differences between the groups was based on one-way ANOVA test, followed by a multiple comparison analysis (Tukey test). The results showed a statistically significant difference between the three groups when using as discriminative features the normalized low frequency (0.04 to 0.15 Hz) HRV (LF-HRV) power and the sample entropy (SE) occurring only between 2:00-3:00, 18:00-19:00, and 19:00-20:00. In addition, the averaged normal-to-normal values show variation during the night time (from 23:00 to 5:00) between the three groups. These results pave the way for further investigation of the interaction of the sympathetic and parasympathetic nervous systems (as reflected in LF - HRV) and the cardiovascular autonomic regulation (as reflected in SE) in LVEF.