Ultra-Wide Band Radar Empowered Driver Drowsiness Detection with Convolutional Spatial Feature Engineering and Artificial Intelligence.

Driving while drowsy poses significant risks, including reduced cognitive function and the potential for accidents, which can lead to severe consequences such as trauma, economic losses, injuries, or death. The use of artificial intelligence can enable effective detection of driver drowsiness, helping to prevent accidents and enhance driver performance. This research aims to address the crucial need for real-time and accurate drowsiness detection to mitigate the impact of fatigue-related accidents. Leveraging ultra-wideband radar data collected over five minutes, the dataset was segmented into one-minute chunks and transformed into grayscale images. Spatial features are retrieved from the images using a two-dimensional Convolutional Neural Network. Following that, these features were used to train and test multiple machine learning classifiers. The ensemble classifier RF-XGB-SVM, which combines Random Forest, XGBoost, and Support Vector Machine using a hard voting criterion, performed admirably with an accuracy of 96.6%. Additionally, the proposed approach was validated with a robust k-fold score of 97% and a standard deviation of 0.018, demonstrating significant results. The dataset is augmented using Generative Adversarial Networks, resulting in improved accuracies for all models. Among them, the RF-XGB-SVM model outperformed the rest with an accuracy score of 99.58%.

A systematic review of physiological signals based driver drowsiness detection systems.

Driving a vehicle is a complex, multidimensional, and potentially risky activity demanding full mobilization and utilization of physiological and cognitive abilities. Drowsiness, often caused by stress, fatigue, and illness declines cognitive capabilities that affect drivers' capability and cause many accidents. Drowsiness-related road accidents are associated with trauma, physical injuries, and fatalities, and often accompany economic loss. Drowsy-related crashes are most common in young people and night shift workers. Real-time and accurate driver drowsiness detection is necessary to bring down the drowsy driving accident rate. Many researchers endeavored for systems to detect drowsiness using different features related to vehicles, and drivers' behavior, as well as, physiological measures. Keeping in view the rising trend in the use of physiological measures, this study presents a comprehensive and systematic review of the recent techniques to detect driver drowsiness using physiological signals. Different sensors augmented with machine learning are utilized which subsequently yield better results. These techniques are analyzed with respect to several aspects such as data collection sensor, environment consideration like controlled or dynamic, experimental set up like real traffic or driving simulators, etc. Similarly, by investigating the type of sensors involved in experiments, this study discusses the advantages and disadvantages of existing studies and points out the research gaps. Perceptions and conceptions are made to provide future research directions for drowsiness detection techniques based on physiological signals.

Empowering Lower Limb Disorder Identification through PoseNet and Artificial Intelligence.

A novel approach is presented in this study for the classification of lower limb disorders, with a specific emphasis on the knee, hip, and ankle. The research employs gait analysis and the extraction of PoseNet features from video data in order to effectively identify and categorize these disorders. The PoseNet algorithm facilitates the extraction of key body joint movements and positions from videos in a non-invasive and user-friendly manner, thereby offering a comprehensive representation of lower limb movements. The features that are extracted are subsequently standardized and employed as inputs for a range of machine learning algorithms, such as Random Forest, Extra Tree Classifier, Multilayer Perceptron, Artificial Neural Networks, and Convolutional Neural Networks. The models undergo training and testing processes using a dataset consisting of 174 real patients and normal individuals collected at the Tehsil Headquarter Hospital Sadiq Abad. The evaluation of their performance is conducted through the utilization of K-fold cross-validation. The findings exhibit a notable level of accuracy and precision in the classification of various lower limb disorders. Notably, the Artificial Neural Networks model achieves the highest accuracy rate of 98.84%. The proposed methodology exhibits potential in enhancing the diagnosis and treatment planning of lower limb disorders. It presents a non-invasive and efficient method of analyzing gait patterns and identifying particular conditions.

Enhancing Diagnosis of Anterior and Inferior Myocardial Infarctions Using UWB Radar and AI-Driven Feature Fusion Approach.

Despite significant improvement in prognosis, myocardial infarction (MI) remains a major cause of morbidity and mortality around the globe. MI is a life-threatening cardiovascular condition that requires prompt diagnosis and appropriate treatment. The primary objective of this research is to identify instances of anterior and inferior myocardial infarction by utilizing data obtained from Ultra-wideband radar technology in a hospital for patients of anterior and inferior MI. The collected data is preprocessed to extract spectral features. A novel feature engineering approach is designed to fuse temporal features and class prediction probability features derived from the spectral feature dataset. Several well-known machine learning models are implemented and fine-tuned to obtain optimal performance in the detection of anterior and inferior MI. The results demonstrate that integration of the fused feature set with machine learning models results in a notable improvement in both the accuracy and precision of MI detection. Notably, random forest (RF) and k-nearest neighbor showed superb performance with an accuracy of 98.8%. For demonstrating the capacity of models to generalize, K-fold cross-validation is carried out, wherein RF exhibits a mean accuracy of 99.1%. Furthermore, the examination of computational complexity indicates a low computational complexity, thereby indicating computational efficiency.

Railway Track Fault Detection Using Selective MFCC Features from Acoustic Data.

Railway track faults may lead to railway accidents and cause human and financial loss. Spatial, temporal, and weather elements, and wear and tear, lead to ballast, loose nuts, misalignment, and cracks leading to accidents. Manual inspection of such defects is time-consuming and prone to errors. Automatic inspection provides a fast, reliable, and unbiased solution. However, highly accurate fault detection is challenging due to the lack of public datasets, noisy data, inefficient models, etc. To obtain better performance, this study presents a novel approach that relies on mel frequency cepstral coefficient features from acoustic data. The primary objective of this study is to increase fault detection performance. As well as designing an ensemble model, we utilize selective features using chi-square(chi2) that have high importance with respect to the target class. Extensive experiments were carried out to analyze the efficiency of the proposed approach. The experimental results suggest that using 60 features, 40 original features, and 20 chi2 features produces optimal results both regarding accuracy and computational complexity. A mean accuracy score of 0.99 was obtained using the proposed approach with machine learning models using the collected data. Moreover, this performance was significantly better than that of existing approaches; however, the performance of models may vary in real-world settings.

Enhancing Cricket Performance Analysis with Human Pose Estimation and Machine Learning.

Cricket has a massive global following and is ranked as the second most popular sport globally, with an estimated 2.5 billion fans. Batting requires quick decisions based on ball speed, trajectory, fielder positions, etc. Recently, computer vision and machine learning techniques have gained attention as potential tools to predict cricket strokes played by batters. This study presents a cutting-edge approach to predicting batsman strokes using computer vision and machine learning. The study analyzes eight strokes: pull, cut, cover drive, straight drive, backfoot punch, on drive, flick, and sweep. The study uses the MediaPipe library to extract features from videos and several machine learning and deep learning algorithms, including random forest (RF), support vector machine, k-nearest neighbors, decision tree, linear regression, and long short-term memory to predict the strokes. The study achieves an outstanding accuracy of 99.77% using the RF algorithm, outperforming the other algorithms used in the study. The k-fold validation of the RF model is 95.0% with a standard deviation of 0.07, highlighting the potential of computer vision and machine learning techniques for predicting batsman strokes in cricket. The study's results could help improve coaching techniques and enhance batsmen's performance in cricket, ultimately improving the game's overall quality.

Predicting bankruptcy of firms using earnings call data and transfer learning.

Business collapse is a common event in economies, small and big alike. A firm's health is crucial to its stakeholders like creditors, investors, partners, etc. and prediction of the upcoming financial crisis is significantly important to devise appropriate strategies to avoid business collapses. Bankruptcy prediction has been regarded as a critical topic in the world of accounting and finance. Methodologies and strategies have been investigated in the research domain for predicting company bankruptcy more promptly and accurately. Conventionally, predicting the financial risk and bankruptcy has been solely achieved using the historic financial data. CEOs also communicate verbally via press releases and voice characteristics, such as emotion and tone may reflect a company's success, according to anecdotal evidence. Companies' publicly available earning calls data is one of the main sources of information to understand how businesses are doing and what are expectations for the next quarters. An earnings call is a conference call between the management of a company and the media. During the call, management offers an overview of recent performance and provides a guide for the next quarter's expectations. The earning calls summary provided by the management can extract CEO's emotions using sentiment analysis. This article investigates the prediction of firms' health in terms of bankruptcy and non-bankruptcy based on emotions extracted from earning calls and proposes a deep learning model in this regard. Features extracted from long short-term memory (LSTM) network are used to train machine learning models. Results show that the models provide results with a high score of 0.93, each for accuracy and F1 when trained on LSTM extracted feature from synthetic minority oversampling technique (SMOTE) balanced data. LSTM features provide better performance than traditional bag of words and TF-IDF features.

Predicting microbe organisms using data of living micro forms of life and hybrid microbes classifier.

Microbe organisms make up approximately 60% of the earth's living matter and the human body is home to millions of microbe organisms. Microbes are microbial threats to health and may lead to several diseases in humans like toxoplasmosis and malaria. The microbiological toxoplasmosis disease in humans is widespread, with a seroprevalence of 3.6-84% in sub-Saharan Africa. This necessitates an automated approach for microbe organisms detection. The primary objective of this study is to predict microbe organisms in the human body. A novel hybrid microbes classifier (HMC) is proposed in this study which is based on a decision tree classifier and extra tree classifier using voting criteria. Experiments involve different machine learning and deep learning models for detecting ten different living microforms of life. Results suggest that the proposed HMC approach achieves a 98% accuracy score, 98% geometric mean score, 97% precision score, and 97% Cohen Kappa score. The proposed model outperforms employed models, as well as, existing state-of-the-art models. Moreover, the k-fold cross-validation corroborates the results as well. The research helps microbiologists identify the type of microbe organisms with high accuracy and prevents many diseases through early detection.

An Approach to Detect Chronic Obstructive Pulmonary Disease Using UWB Radar-Based Temporal and Spectral Features.

Chronic obstructive pulmonary disease (COPD) is a severe and chronic ailment that is currently ranked as the third most common cause of mortality across the globe. COPD patients often experience debilitating symptoms such as chronic coughing, shortness of breath, and fatigue. Sadly, the disease frequently goes undiagnosed until it is too late, leaving patients without the care they desperately need. So, COPD detection at an early stage is crucial to prevent further damage to the lungs and improve quality of life. Traditional COPD detection methods often rely on physical examinations and tests such as spirometry, chest radiography, blood gas tests, and genetic tests. However, these methods may not always be accurate or accessible. One of the key vital signs for detecting COPD is the patient's respiration rate. However, it is crucial to consider a patient's medical and demographic characteristics simultaneously for better detection results. To address this issue, this study aims to detect COPD patients using artificial intelligence techniques. To achieve this goal, a novel framework is proposed that utilizes ultra-wideband (UWB) radar-based temporal and spectral features to build machine learning and deep learning models. This new set of temporal and spectral features is extracted from respiration data collected non-invasively from 1.5 m distance using UWB radar. Different machine learning and deep learning models are trained and tested on the collected dataset. The findings are promising, with a high accuracy score of 100% for COPD detection. This means that the proposed framework could potentially save lives by identifying COPD patients at an early stage. The k-fold cross-validation technique and performance comparison with the state-of-the-art studies are applied to validate its performance, ensuring that the results are robust and reliable. The high accuracy score achieved in the study implies that the proposed framework has the potential for the efficient detection of COPD at an early stage.

Ensemble learning-based feature engineering to analyze maternal health during pregnancy and health risk prediction.

Maternal health is an important aspect of women's health during pregnancy, childbirth, and the postpartum period. Specifically, during pregnancy, different health factors like age, blood disorders, heart rate, etc. can lead to pregnancy complications. Detecting such health factors can alleviate the risk of pregnancy-related complications. This study aims to develop an artificial neural network-based system for predicting maternal health risks using health data records. A novel deep neural network architecture, DT-BiLTCN is proposed that uses decision trees, a bidirectional long short-term memory network, and a temporal convolutional network. Experiments involve using a dataset of 1218 samples collected from maternal health care, hospitals, and community clinics using the IoT-based risk monitoring system. Class imbalance is resolved using the synthetic minority oversampling technique. DT-BiLTCN provides a feature set to obtain high accuracy results which in this case are provided by the support vector machine with a 98% accuracy. Maternal health exploratory data analysis reveals that the health conditions which are the strongest indications of health risk during pregnancy are diastolic and systolic blood pressure, heart rate, and age of pregnant women. Using the proposed model, timely prediction of health risks associated with pregnant women can be made thus mitigating the risk of health complications which helps to save lives.

Automatic User Preferences Selection of Smart Hearing Aid Using BioAid.

Noisy environments, changes and variations in the volume of speech, and non-face-to-face conversations impair the user experience with hearing aids. Generally, a hearing aid amplifies sounds so that a hearing-impaired person can listen, converse, and actively engage in daily activities. Presently, there are some sophisticated hearing aid algorithms available that operate on numerous frequency bands to not only amplify but also provide tuning and noise filtering to minimize background distractions. One of those is the BioAid assistive hearing system, which is an open-source, freely available downloadable app with twenty-four tuning settings. Critically, with this device, a person suffering with hearing loss must manually alter the settings/tuning of their hearing device when their surroundings and scene changes in order to attain a comfortable level of hearing. However, this manual switching among multiple tuning settings is inconvenient and cumbersome since the user is forced to switch to the state that best matches the scene every time the auditory environment changes. The goal of this study is to eliminate this manual switching and automate the BioAid with a scene classification algorithm so that the system automatically identifies the user-selected preferences based on adequate training. The aim of acoustic scene classification is to recognize the audio signature of one of the predefined scene classes that best represent the environment in which it was recorded. BioAid, an open-source biological inspired hearing aid algorithm, is used after conversion to Python. The proposed method consists of two main parts: classification of auditory scenes and selection of hearing aid tuning settings based on user experiences. The DCASE2017 dataset is utilized for scene classification. Among the many classifiers that were trained and tested, random forests have the highest accuracy of 99.7%. In the second part, clean speech audios from the LJ speech dataset are combined with scenes, and the user is asked to listen to the resulting audios and adjust the presets and subsets. A CSV file stores the selection of presets and subsets at which the user can hear clearly against the scenes. Various classifiers are trained on the dataset of user preferences. After training, clean speech audio was convolved with the scene and fed as input to the scene classifier that predicts the scene. The predicted scene was then fed as input to the preset classifier that predicts the user's choice for preset and subset. The BioAid is automatically tuned to the predicted selection. The accuracy of random forest in the prediction of presets and subsets was 100%. This proposed approach has great potential to eliminate the tedious manual switching of hearing assistive device parameters by allowing hearing-impaired individuals to actively participate in daily life by automatically adjusting hearing aid settings based on the acoustic scene.

Railway Track Inspection Using Deep Learning Based on Audio to Spectrogram Conversion: An on-the-Fly Approach.

The periodic inspection of railroad tracks is very important to find structural and geometrical problems that lead to railway accidents. Currently, in Pakistan, rail tracks are inspected by an acoustic-based manual system that requires a railway engineer as a domain expert to differentiate between different rail tracks' faults, which is cumbersome, laborious, and error-prone. This study proposes the use of traditional acoustic-based systems with deep learning models to increase performance and reduce train accidents. Two convolutional neural networks (CNN) models, convolutional 1D and convolutional 2D, and one recurrent neural network (RNN) model, a long short-term memory (LSTM) model, are used in this regard. Initially, three types of faults are considered, including superelevation, wheel burnt, and normal tracks. Contrary to traditional acoustic-based systems where the spectrogram dataset is generated before the model training, the proposed approach uses on-the-fly feature extraction by generating spectrograms as a deep learning model's layer. Different lengths of audio samples are used to analyze their performance with each model. Each audio sample of 17 s is split into 3 variations of 1.7, 3.4, and 8.5 s, and all 3 deep learning models are trained and tested against each split time. Various combinations of audio data augmentation are analyzed extensively to investigate models' performance. The results suggest that the LSTM with 8.5 split time gives the best results with the accuracy of 99.7%, the precision of 99.5%, recall of 99.5%, and F1 score of 99.5%.

Predicting Genetic Disorder and Types of Disorder Using Chain Classifier Approach.

Genetic disorders are the result of mutation in the deoxyribonucleic acid (DNA) sequence which can be developed or inherited from parents. Such mutations may lead to fatal diseases such as Alzheimer's, cancer, Hemochromatosis, etc. Recently, the use of artificial intelligence-based methods has shown superb success in the prediction and prognosis of different diseases. The potential of such methods can be utilized to predict genetic disorders at an early stage using the genome data for timely treatment. This study focuses on the multi-label multi-class problem and makes two major contributions to genetic disorder prediction. A novel feature engineering approach is proposed where the class probabilities from an extra tree (ET) and random forest (RF) are joined to make a feature set for model training. Secondly, the study utilizes the classifier chain approach where multiple classifiers are joined in a chain and the predictions from all the preceding classifiers are used by the conceding classifiers to make the final prediction. Because of the multi-label multi-class data, macro accuracy, Hamming loss, and α-evaluation score are used to evaluate the performance. Results suggest that extreme gradient boosting (XGB) produces the best scores with a 92% α-evaluation score and a 84% macro accuracy score. The performance of XGB is much better than state-of-the-art approaches, in terms of both performance and computational complexity.

Respiration Based Non-Invasive Approach for Emotion Recognition Using Impulse Radio Ultra Wide Band Radar and Machine Learning.

Emotion recognition gained increasingly prominent attraction from a multitude of fields recently due to their wide use in human-computer interaction interface, therapy, and advanced robotics, etc. Human speech, gestures, facial expressions, and physiological signals can be used to recognize different emotions. Despite the discriminating properties to recognize emotions, the first three methods have been regarded as ineffective as the probability of human's voluntary and involuntary concealing the real emotions can not be ignored. Physiological signals, on the other hand, are capable of providing more objective, and reliable emotion recognition. Based on physiological signals, several methods have been introduced for emotion recognition, yet, predominantly such approaches are invasive involving the placement of on-body sensors. The efficacy and accuracy of these approaches are hindered by the sensor malfunctioning and erroneous data due to human limbs movement. This study presents a non-invasive approach where machine learning complements the impulse radio ultra-wideband (IR-UWB) signals for emotion recognition. First, the feasibility of using IR-UWB for emotion recognition is analyzed followed by determining the state of emotions into happiness, disgust, and fear. These emotions are triggered using carefully selected video clips to human subjects involving both males and females. The convincing evidence that different breathing patterns are linked with different emotions has been leveraged to discriminate between different emotions. Chest movement of thirty-five subjects is obtained using IR-UWB radar while watching the video clips in solitude. Extensive signal processing is applied to the obtained chest movement signals to estimate respiration rate per minute (RPM). The RPM estimated by the algorithm is validated by repeated measurements by a commercially available Pulse Oximeter. A dataset is maintained comprising gender, RPM, age, and associated emotions which are further used with several machine learning algorithms for automatic recognition of human emotions. Experiments reveal that IR-UWB possesses the potential to differentiate between different human emotions with a decent accuracy of 76% without placing any on-body sensors. Separate analysis for male and female participants reveals that males experience high arousal for happiness while females experience intense fear emotions. For disgust emotion, no large difference is found for male and female participants. To the best of the authors' knowledge, this study presents the first non-invasive approach using the IR-UWB radar for emotion recognition.

A Novel Approach to Railway Track Faults Detection Using Acoustic Analysis.

Regular inspection of railway track health is crucial for maintaining safe and reliable train operations. Factors, such as cracks, ballast issues, rail discontinuity, loose nuts and bolts, burnt wheels, superelevation, and misalignment developed on the rails due to non-maintenance, pre-emptive investigations and delayed detection, pose a grave danger and threats to the safe operation of rail transport. The traditional procedure of manually inspecting the rail track using a railway cart is both inefficient and prone to human error and biases. In a country like Pakistan where train accidents have taken many lives, it is not unusual to automate such approaches to avoid such accidents and save countless lives. This study aims at enhancing the traditional railway cart system to address these issues by introducing an automatic railway track fault detection system using acoustic analysis. In this regard, this study makes two important contributions: data collection on Pakistan railway tracks using acoustic signals and the application of various classification techniques to the collected data. Initially, three types of tracks are considered, including normal track, wheel burnt and superelevation, due to their common occurrence. Several well-known machine learning algorithms are applied such as support vector machines, logistic regression, random forest and decision tree classifier, in addition to deep learning models like multilayer perceptron and convolutional neural networks. Results suggest that acoustic data can help determine the track faults successfully. Results indicate that the best results are obtained by RF and DT with an accuracy of 97%.

Non-Invasive Driver Drowsiness Detection System.

Drowsiness when in command of a vehicle leads to a decline in cognitive performance that affects driver behavior, potentially causing accidents. Drowsiness-related road accidents lead to severe trauma, economic consequences, impact on others, physical injury and/or even death. Real-time and accurate driver drowsiness detection and warnings systems are necessary schemes to reduce tiredness-related driving accident rates. The research presented here aims at the classification of drowsy and non-drowsy driver states based on respiration rate detection by non-invasive, non-touch, impulsive radio ultra-wideband (IR-UWB) radar. Chest movements of 40 subjects were acquired for 5 m using a lab-placed IR-UWB radar system, and respiration per minute was extracted from the resulting signals. A structured dataset was obtained comprising respiration per minute, age and label (drowsy/non-drowsy). Different machine learning models, namely, Support Vector Machine, Decision Tree, Logistic regression, Gradient Boosting Machine, Extra Tree Classifier and Multilayer Perceptron were trained on the dataset, amongst which the Support Vector Machine shows the best accuracy of 87%. This research provides a ground truth for verification and assessment of UWB to be used effectively for driver drowsiness detection based on respiration.