Narrow Results

Unmanned aerial vehicles (UAVs) can monitor traffic in different scenarios like surveillance, control, and security. The object detection method depends on UAVs equipped with vision sensors, which have received significant attention in domains such as intelligent transportation systems (ITSs) and UAVs, which can monitor road traffic across some distance and offer vital data for following intelligent traffic supervision tasks, namely traffic situational awareness, detecting sudden accidents, and calculating traffic flow. Nevertheless, most vehicle targets exhibit specific features and lesser sizes that challenge accurate vehicle recognition in UAV overhead view. Employing innovative computer vision (CV) models, vehicle recognition and tracking in UAV images contains detecting and following vehicles in aerial footage taken by UAVs. This procedure leverages deep learning (DL) approaches for perfectly detecting vehicles and a robust tracking method for monitoring their actions through the frames, offering vital information for traffic management, surveillance, and urban planning. Therefore, this study designs an Advanced DL-based Vehicle Detection and Tracking on UAV Imagery (ADLVDT-UAVI) approach. The drive of the ADLVDT-UAVI technique is to detect and classify distinct vehicles in the UAV images correctly as Brain-Like Computing technique for Traffic Flow Optimization in Smart Cities. In this approach, Gaussian filtering (GF) primarily eliminates the noise. Besides, the ADLVDT-UAVI technique utilizes a squeeze-and-excitation capsule network (SE-CapsNet) for feature vector derivation. Meanwhile, the hyperparameter selection process involves using the Fractals coati optimization algorithm (COA). Finally, the self-attention bi-directional long short-term memory (SA-BiLSTM) approach is utilized to classify detected vehicles. To validate the improved results of the ADLVDT-UAVI approach, a wide range of experiments is performed under VEDAI and ISPRS Postdam datasets. The experimental validation of the ADLVDT-UAVI approach portrayed the superior accuracy outcome of 98.35% and 98.96% compared to recent models.

The detection of premature ventricular contractions (PVCs) is an effective method for assessing cardiac health and guiding timely interventions to prevent more severe heart conditions. This paper thoroughly investigates the general manifestations of PVCs in electrocardiograms (ECGs) and focuses on three prominent characteristics of PVCs: premature occurrence of the QRS complex, widened QRS complex, and abnormal morphology of the QRS complex. A total of 11 highly interpretable feature parameters were extracted based on these characteristics. Using the MIT-BIH arrhythmia database and leveraging Python libraries NeuroKit and Scikit-learn, a random forest classifier was employed to detect PVCs, achieving an accuracy of 99.55%, a precision of 0.98, a recall of 0.95, and an F1 score of 0.96. The experimental results demonstrate that this detection method not only provides high accuracy with low computational complexity but also offers significant clinical value for effective PVC detection.

By the end of 2023, China’s population aged 60 and above is projected to reach 296.97 million, or 21.1% of the total population, with those aged 65 and above numbering 216.76 million, accounting for 15.4%. This marks China’s transition into a moderately aged society. Addressing the diverse needs of the elderly and solving the social issues arising from population aging are crucial for the nation’s overall development and public welfare. A comprehensive analysis of the living environment of the elderly can aid in enhancing care services for this group. This study analyzes data from home-tested and hospitalized elderly patients to explore differences across various indicators. It classifies patients based on factors such as stress test comprehensive index, self-rating stress scale (SDS score), age, and gender, evaluating model performance using metrics like accuracy and recall. Principal component analysis (PCA) and K-means clustering have revealed key differences between the two groups, while Random Forest, Support Vector Machine (SVM), and XGBoost were employed for classification modeling to predict whether elderly patients are at home or in hospital. The results indicate significant differences in the feature space between the two patient groups, with feature analysis accurately predicting patient classification. This study offers valuable data support for elderly health management, suggesting that future research should consider more complex models or additional features to improve prediction accuracy and robustness. The findings highlight the importance of tailored care strategies to address the diverse needs of an aging population, ultimately contributing to better healthcare outcomes for the elderly.

Acute respiratory distress syndrome (ARDS) is a critical condition that causes alveolar injury and impairs gas exchange. Proper setting of the initial tidal volume during mechanical ventilation is crucial for patient survival and recovery. This study aims to predict the initial tidal volume for ARDS patients using a multiple linear regression model based on data from 1580 Medical Information Mart for Intensive Care IV (MIMIC-IV) and 2295 eICU Collaborative Research Database (eICU-CRD) patients receiving noninvasive respiratory support. The model incorporates 15 variables, including gender, age, pH, and blood gas indicators. Regression coefficients were estimated using the least squares method, and model performance was evaluated using metrics such as mean squared error (MSE), R-squared ($R^2$), mean absolute error (MAE), and mean absolute percentage error (MAPE). The results show that the model has good stability and generalization, with the best performance achieving an MSE of 824.89 and $R^2$ of 0.73. These findings suggest a significant correlation between the selected features and tidal volume, providing a basis for personalized treatment and aiding clinicians in mechanical ventilation decision-making for ARDS patients.

The research aimed to develop an effective prognostic model for osteosarcoma. This study included 72 osteosarcoma cases from the Target OS database and 87 cases from the Shanghai General Hospital (SGH) Cohort. DNA methylation data and RNA-seq data collected from the Target OS database and the SGH Cohort were processed using synthetic minority over-sampling technique (SMOTE), Tomek links, convolutional neural network (CNN), and Random Forest models. Based on the prognosis of the patients, we divided the cases into a death/survival group and a recurrence/nonrecurrence group. A total of 10 times of model training and testing were performed separately, using the test results as model evaluation parameters. We evaluated the model performance by comprehensively assessing its precision, recall, accuracy, and F1 score. The results of this study offer a robust methodology for determining the prognosis and characteristics of osteosarcoma patients, applicable even to smaller data sets. This approach provides novel insights into the molecular data analysis of rare diseases.

Watching sports matches is a beloved pastime for every fan, and predicting the results and betting on which team will win adds to the excitement. Unfortunately, for some users, a lack of self-control can turn initially innocent play into a long-term problem. For this reason, problem gambling is among the crucial issues facing modern societies. Bookmaker companies, therefore, put considerable effort into implementing responsible gambling practices; however, identifying at-risk individuals remains a major challenge. Our paper focuses on identifying users who may currently have, or soon develop, potential issues with gambling. To achieve this, we introduce a set of machine learning methods combined with preprocessing tools that allow us to initially acquire and anonymize user data. This anonymized database is then used to identify high-risk groups. We tested our approach on a large dataset that was obtained and preprocessed specifically for this study. The experiments were conducted on actual data and verified by specialists responsible for identifying gambling problems. Using our method, we successfully identified and detected the early signs of potential gambling problems in multiple users.

Accurate predictions of sound radiation are crucial for assessing sound emissions in the far field. A widely used approach is the boundary element method, which traditionally solves the Kirchhoff-Helmholtz integral equation by discretization. While the boundary integral formulation inherently satisfies the Sommerfeld radiation condition, a significant drawback of the boundary element method is its difficulty in incorporating noisy measurement data. In recent years, physics-informed machine learning approaches have demonstrated robust predictions of physical problems, even in the presence of noisy or imperfect data. To utilize these benefits while addressing acoustic predictions in unbounded domains, this study employs boundary integral neural networks for predicting acoustic radiation. These networks incorporate the residual of the boundary integral equation into the neural network’s loss function, enabling data-driven predictions of acoustic radiation from noisy boundary data. The results demonstrate that boundary integral neural networks are able to accurately predict the sound pressure field for both interior and exterior problems of a two-dimensional acoustic domain. The study also highlights that the data-driven approach outperforms the conventional boundary element method, particularly at high noise levels. Consequently, the presented method offers a promising approach for predicting sound radiation based on noisy surface vibration measurements.

This study introduced a new approach for monitoring regional development by applying satellite data with machine learning algorithms. Satellite data that represent physical features and environmental factors were obtained by developing a web-based application on the Google Earth Engine platform. Four machine learning methods were applied to the obtained geospatial data to predict provincial gross domestic product. The random forest method achieved the highest predictive performance, with 97.7% accuracy. The constructed random forest model was extended to conduct variable importance and minimal depth analyses, enabling the quantification of a factor’s influence on the prediction outcome. Variable importance and minimal depth analyses generated similar results, indicating that urban area and population are the most influential factors. Moreover, environmental and climate indicators exert medium-level effects. This study showed that integrating available satellite data and machine learning methods could be an alternative framework for facilitating a timely and costless monitoring system of regional development.

Detailed data on the distribution of human populations are valuable inputs to research and decision making. This study aims at compiling data on population density that are more granular than government-published estimates and assessing different methods and model specifications. As a first step, we combine government-published data with publicly available data like land cover classes, elevation, slope, and nighttime lights, and then apply a random forest approach to estimate population density in the Philippines and Thailand at the 100 meter (m) by 100m level. Second, we use different specifications of random forest and Bayesian model averaging (BMA) techniques to forecast grid-level population density and evaluate their predictive power. The use of a random forest model showed that reasonable forecasts of grid-level population growth rates are achievable. The results of this study contribute to the assessment of methods like random forest and BMA in forecasting population distributions.

Parkinsonism is a clinical syndrome characterized by the progressive loss of striatal dopamine. Its diagnosis is usually corroborated by neuroimaging data such as DaTSCAN neuroimages that allow visualizing the possible dopamine deficiency. During the last decade, a number of computer systems have been proposed to automatically analyze DaTSCAN neuroimages, eliminating the subjectivity inherent to the visual examination of the data. In this work, we propose a computer system based on machine learning to separate Parkinsonian patients and control subjects using the size and shape of the striatal region, modeled from DaTSCAN data. First, an algorithm based on adaptative thresholding is used to parcel the striatum. This region is then divided into two according to the brain hemisphere division and characterized with 152 measures, extracted from the volume and its three possible 2-dimensional projections. Afterwards, the Bhattacharyya distance is used to discard the least discriminative measures and, finally, the neuroimage category is estimated by means of a Support Vector Machine classifier. This method was evaluated using a dataset with 189 DaTSCAN neuroimages, obtaining an accuracy rate over 94%. This rate outperforms those obtained by previous approaches that use the intensity of each striatal voxel as a feature.

Machine learning (ML) systems are affected by a pervasive lack of transparency. The eXplainable Artificial Intelligence (XAI) research area addresses this problem and the related issue of explaining the behavior of ML systems in terms that are understandable to human beings. In many explanation of XAI approaches, the output of ML systems are explained in terms of low-level features of their inputs. However, these approaches leave a substantive explanatory burden with human users, insofar as the latter are required to map low-level properties into more salient and readily understandable parts of the input. To alleviate this cognitive burden, an alternative model-agnostic framework is proposed here. This framework is instantiated to address explanation problems in the context of ML image classification systems, without relying on pixel relevance maps and other low-level features of the input. More specifically, one obtains sets of middle-level properties of classification inputs that are perceptually salient by applying sparse dictionary learning techniques. These middle-level properties are used as building blocks for explanations of image classifications. The achieved explanations are parsimonious, for their reliance on a limited set of middle-level image properties. And they can be contrastive, because the set of middle-level image properties can be used to explain why the system advanced the proposed classification over other antagonist classifications. In view of its model-agnostic character, the proposed framework is adaptable to a variety of other ML systems and explanation problems.

Autism Spectrum Disorder (ASD) is a largely prevalent neurodevelopmental condition with a big social and economical impact affecting the entire life of families. There is an intense search for biomarkers that can be assessed as early as possible in order to initiate treatment and preparation of the family to deal with the challenges imposed by the condition. Brain imaging biomarkers have special interest. Specifically, functional connectivity data extracted from resting state functional magnetic resonance imaging (rs-fMRI) should allow to detect brain connectivity alterations. Machine learning pipelines encompass the estimation of the functional connectivity matrix from brain parcellations, feature extraction, and building classification models for ASD prediction. The works reported in the literature are very heterogeneous from the computational and methodological point of view. In this paper, we carry out a comprehensive computational exploration of the impact of the choices involved while building these machine learning pipelines. Specifically, we consider six brain parcellation definitions, five methods for functional connectivity matrix construction, six feature extraction/selection approaches, and nine classifier building algorithms. We report the prediction performance sensitivity to each of these choices, as well as the best results that are comparable with the state of the art.

Over the last decades, the exuberant development of next-generation sequencing has revolutionized gene discovery. These technologies have boosted the mapping of single nucleotide polymorphisms (SNPs) across the human genome, providing a complex universe of heterogeneity characterizing individuals worldwide. Fractal dimension (FD) measures the degree of geometric irregularity, quantifying how “complex” a self-similar natural phenomenon is. We compared two FD algorithms, box-counting dimension (BCD) and Higuchi’s fractal dimension (HFD), to characterize genome-wide patterns of SNPs extracted from the HapMap data set, which includes data from 1184 healthy subjects of eleven populations. In addition, we have used cluster and classification analysis to relate the genetic distances within chromosomes based on FD similarities to the geographical distances among the 11 global populations. We found that HFD outperformed BCD at both grand average clusterization analysis by the cophenetic correlation coefficient, in which the closest value to 1 represents the most accurate clustering solution (0.981 for the HFD and 0.956 for the BCD) and classification (79.0% accuracy, 61.7% sensitivity, and 96.4% specificity for the HFD with respect to 69.1% accuracy, 43.2% sensitivity, and 94.9% specificity for the BCD) of the 11 populations present in the HapMap data set. These results support the evidence that HFD is a reliable measure helpful in representing individual variations within all chromosomes and categorizing individuals and global populations.

Parkinson’s Disease (PD) is the second most prevalent neurodegenerative disorder among adults. Although its triggers are still not clear, they may be due to a combination of different types of biomarkers measured through medical imaging, metabolomics, proteomics or genetics, among others. In this context, we have proposed a Computer-Aided Diagnosis (CAD) system that combines structural and functional imaging data from subjects in Parkinson’s Progression Markers Initiative dataset by means of an Ensemble Learning methodology trained to identify and penalize input sources with low classification rates and/ or high-variability. This proposal improves results published in recent years and provides an accurate solution not only from the point of view of image preprocessing (including a comparison between different intensity preservation techniques), but also in terms of dimensionality reduction methods (Isomap). In addition, we have also introduced a bagging classification schema for scenarios with unbalanced data.

As shown by our results, the CAD proposal is able to detect PD with $96.48\%$ of balanced accuracy, and opens up the possibility of combining any number of input data sources relevant for PD.

Spiking Neural Networks (SNNs) help achieve brain-like efficiency and functionality by building neurons and synapses that mimic the human brain’s transmission of electrical signals. However, optimal SNN implementation requires a precise balance of parametric values. To design such ubiquitous neural networks, a graphical tool for visualizing, analyzing, and explaining the internal behavior of spikes is crucial. Although some popular SNN simulators are available, these tools do not allow users to interact with the neural network during simulation. To this end, we have introduced the first runtime interactive simulator, called Runtime Analyzing and Visualization Simulator (RAVSim),^a developed to analyze and dynamically visualize the behavior of SNNs, allowing end-users to interact, observe output concentration reactions, and make changes directly during the simulation. In this paper, we present RAVSim with the current implementation of runtime interaction using the LIF neural model with different connectivity schemes, an image classification model using SNNs, and a dataset creation feature. Our main objective is to primarily investigate binary classification using SNNs with RGB images. We created a feed-forward network using the LIF neural model for an image classification algorithm and evaluated it by using RAVSim. The algorithm classifies faces with and without masks, achieving an accuracy of 91.8% using 1000 neurons in a hidden layer, 0.0758 MSE, and an execution time of ∼10min on the CPU. The experimental results show that using RAVSim not only increases network design speed but also accelerates user learning capability.

Brain–computer interfaces (BCIs) provide communicative alternatives to those without functional speech. Covert speech (CS)-based BCIs enable communication simply by thinking of words and thus have intuitive appeal. However, an elusive barrier to their clinical translation is the collection of voluminous examples of high-quality CS signals, as iteratively rehearsing words for long durations is mentally fatiguing. Research on CS and speech perception (SP) identifies common spatiotemporal patterns in their respective electroencephalographic (EEG) signals, pointing towards shared encoding mechanisms. The goal of this study was to investigate whether a model that leverages the signal similarities between SP and CS can differentiate speech-related EEG signals online. Ten participants completed a dyadic protocol where in each trial, they listened to a randomly selected word and then subsequently mentally rehearsed the word. In the offline sessions, eight words were presented to participants. For the subsequent online sessions, the two most distinct words (most separable in terms of their EEG signals) were chosen to form a ternary classification problem (two words and rest). The model comprised a functional mapping derived from SP and CS signals of the same speech token (features are extracted via a Riemannian approach). An average ternary online accuracy of 75.3% (60% chance level) was achieved across participants, with individual accuracies as high as 93%. Moreover, we observed that the signal-to-noise ratio (SNR) of CS signals was enhanced by perception-covert modeling according to the level of high-frequency ($\gamma$-band) correspondence between CS and SP. These findings may lead to less burdensome data collection for training speech BCIs, which could eventually enhance the rate at which the vocabulary can grow.

Artificial intelligence (AI)-based approaches are crucial in computer-aided diagnosis (CAD) for various medical applications. Their ability to quickly and accurately learn from complex data is remarkable. Deep learning (DL) models have shown promising results in accurately classifying Alzheimer’s disease (AD) and its related cognitive states, Early Mild Cognitive Impairment (EMCI) and Late Mild Cognitive Impairment (LMCI), along with the healthy conditions known as Cognitively Normal (CN). This offers valuable insights into disease progression and diagnosis. However, certain traditional machine learning (ML) classifiers perform equally well or even better than DL models, requiring less training data. This is particularly valuable in CAD in situations with limited labeled datasets. In this paper, we propose an ensemble classifier based on ML models for magnetic resonance imaging (MRI) data, which achieved an impressive accuracy of 96.52%. This represents a 3–5% improvement over the best individual classifier. We evaluated popular ML classifiers for AD classification under both data-scarce and data-rich conditions using the Alzheimer’s Disease Neuroimaging Initiative and Open Access Series of Imaging Studies datasets. By comparing the results to state-of-the-art CNN-centric DL algorithms, we gain insights into the strengths and weaknesses of each approach. This work will help users to select the most suitable algorithm for AD classification based on data availability.

Eye movements are the primary way primates interact with the world. Understanding how the brain controls the eyes is therefore crucial for improving human health and designing visual rehabilitation devices. However, brain activity is challenging to decipher. Here, we leveraged machine learning algorithms to reconstruct tracking eye movements from high-resolution neuronal recordings. We found that continuous eye position could be decoded with high accuracy using spiking data from only a few dozen cortical neurons. We tested eight decoders and found that neural network models yielded the highest decoding accuracy. Simpler models performed well above chance with a substantial reduction in training time. We measured the impact of data quantity (e.g. number of neurons) and data format (e.g. bin width) on training time, inference time, and generalizability. Training models with more input data improved performance, as expected, but the format of the behavioral output was critical for emphasizing or omitting specific oculomotor events. Our results provide the first demonstration, to our knowledge, of continuously decoded eye movements across a large field of view. Our comprehensive investigation of predictive power and computational efficiency for common decoder architectures provides a much-needed foundation for future work on real-time gaze-tracking devices.

Applications of Artificial Intelligence (AI) are revolutionizing biomedical research and healthcare by offering data-driven predictions that assist in diagnoses. Supervised learning systems are trained on large datasets to predict outcomes for new test cases. However, they typically do not provide an indication of the reliability of these predictions, even though error estimates are integral to model development. Here, we introduce a novel method to identify regions in the feature space that diverge from training data, where an AI model may perform poorly. We utilize a compact precompiled structure that allows for fast and direct access to confidence scores in real time at the point of use without requiring access to the training data or model algorithms. As a result, users can determine when to trust the AI model’s outputs, while developers can identify where the model’s applicability is limited. We validate our approach using simulated data and several biomedical case studies, demonstrating that our approach provides fast confidence estimates ($<0.2$ milliseconds per case), with high concordance to previously developed methods (f-$\mathrm{\text{score}}>0.965$). These estimates can be easily added to real-world AI applications. We argue that providing confidence estimates should be a standard practice for all AI applications in public use.

Machine learning has blossomed in recent decades and has become essential in many fields. It significantly solved some problems in particle physics — particle reconstruction, event classification, etc. However, it is now time to break the limitation of conventional machine learning with quantum computing. A support-vector machine algorithm with a quantum kernel estimator (QSVM-Kernel) leverages high-dimensional quantum state space to identify a signal from backgrounds. In this study, we have pioneered employing this quantum machine learning algorithm to study the $e^{+}{e}^{-}\to ZH$ process at the Circular Electron–Positron Collider (CEPC), a proposed Higgs factory to study electroweak symmetry breaking of particle physics. Using 6 qubits on quantum computer simulators, we optimized the QSVM-Kernel algorithm and obtained a classification performance similar to the classical support-vector machine algorithm. Furthermore, we have validated the QSVM-Kernel algorithm using 6-qubits on quantum computer hardware from both IBM and Origin Quantum: the classification performances of both are approaching noiseless quantum computer simulators. In addition, the Origin Quantum hardware results are similar to the IBM Quantum hardware within the uncertainties in our study. Our study shows that state-of-the-art quantum computing technologies could be utilized by particle physics, a branch of fundamental science that relies on big experimental data.