Abstract: This disclosure relates generally to methods and systems for unobtrusive digital health assessment of high risk subjects, wherein bio-markers pertaining to a disease are identified automatically using physical activity and physiology monitoring on a continuous basis. Identification of bio-markers in the medical domain is conventionally dependent on insights derived from medical tests which are obtrusive in nature. Systems and methods of the present disclosure integrate physical characteristics, lifestyle habits and prevailing medical conditions with monitored physical activities and physiological measurements to assess health of high risk subjects. Systems and methods of the present disclosure also enable automatic generation of control class and treatment class that may be effectively used for health assessment.

Abstract: Conventionally, Atrial Fibrillation (AF) has been detected using atrial analyses which is vulnerable to background noise. Again there is a dependency on statistical features which are extracted from R-R intervals of long ECG recordings. The present disclosure addresses AF detection from single lead short ECG recordings of less than one minute wherein automatic detection of P-R and P-Q intervals is difficult, which introduces error in feature computing from the segregated intervals and compromises the performance of the classifier. In the present disclosure, a Recurrent Neural Network (RNN) based architecture comprising two Long Short Term Memory (LSTM) networks is provided for temporal analysis of R-R intervals and P wave regions in an ECG signal respectively. Output sates of the two LSTM networks are merged at a dense layer along with a set of hand-crafted statistical features to create a composite feature set for classification of the AF.

Abstract: Computer-aided diagnosis algorithms require a large volume of training data. The existing methods for simulating artificial biomedical signals are mostly based on physics driven mathematical models that require too many assumptions, making them challenging to simulate on a large scale. Alternatively, conventional deep learning-based approaches are pure data driven and hence, do not have physiological interpretation. The present disclosure provides a method that effectively combines both physiological domain knowledge and deep learning to enable simulation of realistic cardiovascular disease specific biomedical signals. An ensemble Generative Adversarial Network (GAN) including a Long Short-Term Memory GAN (LSTM-GAN) configured to generate a Heart Rate Variability (HRV) pattern associated with the cardiovascular disease condition and a Deep Convolutional GAN (DCGAN) configured to create a morphology of a representative cardiac cycle is provided.

Abstract: A method and system for a feature subset-classifier pair for a classification task. The classification task corresponds to automatically classifying data associated with a subject(s) or object(s) of interest into an appropriate class based on a feature subset selected among a plurality of features extracted from the data and a classifier selected from a set of classifier types. The method proposed includes simultaneously determining the feature subset-classifier pair based on a relax-greedy {feature subset, classifier} approach utilizing sub-greedy search process based on a patience function, wherein the feature subset-classifier pair provides an optimal combination for more accurate classification. The automatic joint selection is time efficient solution, effectively speeding up the classification task.

Abstract: Conventional systems and methods of classifying heart signals include segmenting them which can fail due to the presence of noise, artifacts and other sounds including third heart sound ‘S3’, fourth heart sound ‘S4’, and murmur. Heart sounds are inherently prone to interfering noise (ambient, speech, etc.) and motion artifact, which can overlap time location and frequency spectra of murmur in heart sound. Embodiments of the present disclosure provide parallel implementation of Deep Neural Networks (DNN) for classifying heart sound signals (HSS) wherein spatial (presence of different frequencies component) filters from Spectrogram feature(s) of the HSS are learnt by a first DNN while time-varying component of the signals from MFCC features of the HSS are learnt by a second DNN for classifying the heart sound signal as one of normal sound signal or murmur sound signal.

Abstract: Monitoring the quality of sleep of an individual is essential for ensuring one's overall well-being. The existing methods for non-apnea sleep arousal detection are manual. A system and method for the non-apnea sleep arousal detection has been provided. The method uses a feature engineering based binary classification approach for distinguishing non-apnea arousal and non-arousal. A training data set is prepared using a plurality of physiological signals. A plurality of features are derived from the training data set. Out of those only a set of features are selected for training a plurality of random forest classifier models. A test sample is then provided to the plurality of random forest classifier models in the instances of fixed duration. This results in generation of prediction probabilities for each instances. The prediction probabilities are then used to predict the probabilities of non-apnea sleep arousal in the test sample.

Abstract: A system and method for classifying the phonocardiogram (PCG) signal quality has been described. The system is configured to identify the quality of the PCG signal recording and accepting only diagnosable quality recordings for further cardiac analysis. The system includes the derivation of plurality features of the PCG signal from the training dataset. The extracted features are preprocessed and are then ranked using mRMR algorithm. Based on the ranking the irrelevant and redundant features are rejected if their mRMR strength is less. A training model is generated using the relevant set of features. The PCG signal of the person under test is captured using a digital stethoscope and a smartphone. The PCG signal is preprocessed and only the relevant set of features are extracted. And finally the PCG signal is classified into diagnosable or non-diagnosable using the relevant set of features and a random forest classifier.

Abstract: A method for measuring blood pressure of a subject is described herein. In an implementation, the method includes obtaining a plurality of photoplethysmogram (PPG) features associated with the subject. The method further includes ascertaining one or more latent parameters associated with the subject based on the plurality of PPG features and a reference model, wherein the reference model indicates a correlation between the plurality of PPG features and the one or more latent parameters. Further, blood pressure of the subject is determined based on the one or more latent parameters and the plurality of PPG features.

Abstract: A system and method for estimating blood pressure (BP) using photoplethysmogram (PPG) has been explained. The PPG is captured from a PPG sensor (102). For preparing training model, a pulse oximeter is used for capturing PPG. For testing, a smartphone camera is used for capturing the PPG signal. A plurality of features are extracted from the preprocessed PPG signal. A BP distribution is then generated using the plurality of features and the training model. The BP distribution is part of a set of BP distributions generated from different subjects. Finally, a post-processing methodology have been used to reject inconsistent data out of the set of BP distributions and BP value is estimated only for the remaining BP distributions and a statistical average is provided as the blood pressure estimate.

Abstract: Non-invasive methods for accurately classifying Coronary Artery Disease (CAD) is a challenging task. In the present disclosure, a two stage classification is performed. In the first stage of classification, a metadata based rule engine is utilized to classify a subject into one of a confirmed CAD subject, a CAD subject and a non-CAD subject. Here, a set of optimal parameters are selected from a set of metadata associated with the subject based on a difference in frequency of occurrence of the CAD among a disease population and a non-disease population. Further, an optimal threshold associated with each optimal parameter is calculated based on an inflexion based correlation analysis. Further, the CAD subject, classified by the metadata based rule engine is further reclassified in a second stage by utilizing a set of cardiovascular signal into one of the CAD subject and the non-CAD subject.

Abstract: The disclosure generally relates to methods and systems for identifying presence of abnormal heart sounds from heart sound signals of a subject being monitored. Conventional Artificial intelligence (AI) based abnormal heart sounds detection models with supervised learning requires a substantial amount of accurate training datasets covering all heart disease types for the training, which is quiet challenging. The present methods and systems solve the problem solves the problem of identifying presence of the abnormal heart sounds using an efficient semi-supervised learning model. The semi-supervised learning model is generated based on probability distribution of spectrographic properties obtained from heart sound signals of healthy subjects. A Kullback-Leibler (KL) divergence between a predefined Gaussian distribution and an encoded probability distribution of the semi-supervised learning model is determined as an anomaly score for identifying the abnormal heart sounds.

Abstract: A method and system for detection of coronary artery disease (CAD) in a person using a fusion approach has been described. The invention the detection of CAD in the person by capturing of a plurality of physiological signals such as phonocardiogram (PCG), photoplethysmograph (PPG), ECG, galvanic skin response (GSR) etc. from the person. A plurality of features are extracted from the physiological signals. The person is then classified as CAD or normal using the each of the features independently. The classification is done based on supervised machine learning technique. The output of the classification is then fused and used for the detection of the CAD in the person using a predefined criteria.

Abstract: Electrocardiography (ECG) signals contain important markers for Coronary Heart Disease (CHD). State of the art systems and methods rely on clinically available multi-lead ECG for CHD classification which is not cost effective. Moreover the state of the art methods are applied on digital ECG time series data only. Also, discriminative HRV markers are not often present in short ECG recordings necessitating long hours of ECG data to analyze. In accordance with the present disclosure, systems and methods described hereinafter extract ECG time series from ECG images obtained from commercially available low-cost single lead ECG devices through a combination of image and signal processing steps including Histogram analysis, Morphological operation-thinning, Extraction of lines, Extraction of Reference Pulse, Extraction of ECG and interpolating missing data.

Abstract: A method and system for blood pressure (BP) estimation of a person is provided. The system is estimating pulse transit time (PTT) using the ECG signal and PPG signal of the person. A plurality of features are extracted from the PPG. The plurality of PPG features and the PTT are provided as inputs to an automated feature selection algorithm. This algorithm selects a set of features suitable for BP estimation. The selected features are fed to a classifier to classify the database into low/normal BP range and a high BP range. The correctly classified normal BP data are then used to create a regression model to predict BP from the selected features. The current methodology uses automated feature selection mechanism and also employs a block to reject extreme BP data. Thus the available accuracy in predicting BP is expected to be more than the existing BP estimation methods.

Abstract: Embodiments herein provide a system and method for screening and monitoring of cardiac diseases by analyzing acquired physiological signals. Unlike state of art approaches that consider only synchronized ECG and PPG signals for cardiac health analysis and do not consider PCG which is a critical signal for CAD analysis, the system synchronously captures physiological signals such as photo plethysmograph (PPG), phonocardiogram (PCG) and electrocardiogram (ECG) from subject(s) and builds an analytical model in the cloud for analyzing heart conditions from the captured physiological signals. The system and method provides a fusion based approach of combining the captured physiological signals such as PPG, PCG and ECG along with other details such as subject clinical information, demography information and so on. The analytical model is pretrained using ECG. PPG and PCG along with metadata associated with the subject such as demography and clinical information.

Abstract: Conventionally, Atrial Fibrillation (AF) has been detected using atrial analyses which is vulnerable to background noise. Again there is a dependency on statistical features which are extracted from R-R intervals of long ECG recordings. The present disclosure addresses AF detection from single lead short ECG recordings of less than one minute wherein automatic detection of P-R and P-Q intervals is difficult, which introduces error in feature computing from the segregated intervals and compromises the performance of the classifier. In the present disclosure, a Recurrent Neural Network (RNN) based architecture comprising two Long Short Term Memory (LSTM) networks is provided for temporal analysis of R-R intervals and P wave regions in an ECG signal respectively. Output sates of the two LSTM networks are merged at a dense layer along with a set of hand-crafted statistical features to create a composite feature set for classification of the AF.

Abstract: Current technologies analyze electrocardiogram (ECG) signals for a long duration, which is not always a practical scenario. Moreover the current scenarios perform a binary classification between normal and Atrial Fibrillation (AF) only, whereas there are many abnormal rhythms apart from AF. Conventional systems/methods have their own limitations and may tend to misclassify ECG signals, thereby resulting in an unbalanced multi-label classification problem.

Abstract: Monitoring the quality of sleep of an individual is essential for ensuring one's overall well-being. The existing methods for non-apnea sleep arousal detection are manual. A system and method for the non-apnea sleep arousal detection has been provided. The method uses a feature engineering based binary classification approach for distinguishing non-apnea arousal and non-arousal. A training data set is prepared using a plurality of physiological signals. A plurality of features are derived from the training data set. Out of those only a set of features are selected for training a plurality of random forest classifier models. A test sample is then provided to the plurality of random forest classifier models in the instances of fixed duration. This results in generation of prediction probabilities for each instances. The prediction probabilities are then used to predict the probabilities of non-apnea sleep arousal in the test sample.

Abstract: Electrocardiography (ECG) signals contain important markers for Coronary Heart Disease (CHD). State of the art systems and methods rely on clinically available multi-lead ECG for CHD classification which is not cost effective. Moreover the state of the art methods are applied on digital ECG time series data only. Also, discriminative HRV markers are not often present in short ECG recordings necessitating long hours of ECG data to analyze. In accordance with the present disclosure, systems and methods described hereinafter extract ECG time series from ECG images obtained from commercially available low-cost single lead ECG devices through a combination of image and signal processing steps including Histogram analysis, Morphological operation-thinning, Extraction of lines, Extraction of Reference Pulse, Extraction of ECG and interpolating missing data.

Abstract: Conventional systems and methods of classifying heart signals include segmenting them which can fail due to the presence of noise, artifacts and other sounds including third heart sound ‘S3’, fourth heart sound ‘S4’, and murmur. Heart sounds are inherently prone to interfering noise (ambient, speech, etc.) and motion artifact, which can overlap time location and frequency spectra of murmur in heart sound. Embodiments of the present disclosure provide parallel implementation of Deep Neural Networks (DNN) for classifying heart sound signals (HSS) wherein spatial (presence of different frequencies component) filters from Spectrogram feature(s) of the HSS are learnt by a first DNN while time-varying component of the signals from MFCC features of the HSS are learnt by a second DNN for classifying the heart sound signal as one of normal sound signal or murmur sound signal.