Over the past two decades, machine learning systems have been proliferating in the healthcare industry domains, such as digital health, fitness tracking, patient monitoring, and disease diagnostics. In addition to this, with technological advancement, physiological sensors paired with artificial intelligence have acquired people’s attention because of their multifarious advantages. Such sensors are predominantly inexpensive, portable, easy to use and can help measure health parameters continuously and non-invasively using artificial intelligence. Technologies, such as PPG (Photoplethysmography) and ECG (Electrocardiography), are two promising techniques with immense potential that can track cardiovascular health with significant impact. In this survey paper, we aim to analyse, summarise, and compare the state-of-the-art methods for machine learning-based blood pressure estimation in a continuous, cuffless, and non-invasive manner by PPG biosignals. This survey divides the research work into two machine learning categories: shallow learning and deep learning. PPG feature extraction techniques and datasets are also presented in this paper. Additionally, a concise comparative analysis of PPG and ECG has been provided from the literature. Moreover, to compare different state-of-the-art traditional feature extraction techniques using PPG biosignals, we applied several machine learning algorithms to predict hypertension and heart rate estimation. Finally, we conclude by summarising critical implications and propose some promising future perspectives that will lead to advancements in this domain. 

A survey: From shallow to deep machine learning approaches for blood pressure estimation using biosensors

Photoplethysmography (PPG) sensor-enabled wearable health monitoring devices can monitor realtime health status. PPG technology is a low-cost, noninvasive optical method used to measure a volumetric change in blood during a cardiac cycle. Continues analysis of change in light signal due to change in the blood helps medical professionals to extract valuable information regarding the cardiovascular system. Traditionally, an electrocardiogram (ECG) has been used as a dominant monitoring technique to detect irregularities in the cardiovascular system. However, in ECG for monitoring cardiac status, several electrodes have to be placed at different body locations, limiting its uses under medical assistantship and in a stationary position. Therefore, to fulfill the market demand for wearable and portable health monitoring devices, researchers are now showing interest in the PPG sensor enable wearable devices. However, the robustness of PPG sensor-enabled wearable devices is highly deviating due to motion artifacts. Therefore before extracting vital sign information like heart rate with PPG sensor, efficient removal of motion artifact is very important. This review orients the research survey on the principles and methods proposed for denoising and heart rate peak detection with PPG. The efficacy of each method related to heart rate peak detection with PPG technologies was compared in terms of mean absolute error, error percentage, and correlation coefficient. A comparative analysis is formulated to estimate heart rate based on the literature survey from the last ten years on PPG technology. This review article aims to explore different methods and challenges mentioned in state-of-the-art research related to motion artifacts removal and heart rate estimation from PPG-enabled wearable devices.

A Review on Computation Methods Used in Photoplethysmography Signal Analysis for Heart Rate Estimation

Cardiac disorders are one of the prime reasons for an increasing global death rate. Reliable and efficient diagnosis procedures are imperative to minimize the risk posed by heart disorders. Computer‐aided diagnosis, based on machine learning and biomedical signal analysis, has recently been adopted by researchers to accurately predict cardiac ailments. Multi‐channel Electrocardiogram signals are mostly used in scientific literature as an indicator to diagnose cardiac disorders. Recently pulse plethysmograph (PuPG) signal got attention as an evolving biosignal and promising diagnostic tool to detect heart disorders since it has a simple sensor with low cost, non‐invasive, reliable, and easy to handle technology. This article proposes a computer‐aided diagnosis system to detect Myocardial Infarction, Dilated Cardiomyopathy, and Hypertension from PuPG signals. Raw PuPG signal is first preprocessed through empirical mode decomposition (EMD) by removing the redundant and useless information content. Then, highly discriminative features are extracted from preprocessed PuPG signal through novel local spectral ternary patterns (LSTP). Extracted LSTPs are input to a variety of classification methods such as support vector machines (SVM), K‐nearest neighbours, decision tree, and so on. SVM with cubic kernel yielded the best classification performance of 98.4% accuracy, 96.7% sensitivity, and 99.6% specificity with 10‐fold cross‐validation. The proposed framework was trained and tested on a self‐collected PuPG signals database of heart disorders. A comparison with previous studies and other feature descriptors shows the superiority of the proposed system. This research provides better insights into the contributions of PuPG signals towards reliable detection of heart disorder through low‐cost and non‐invasive means.

Pulse plethysmograph signal analysis method for classification of heart diseases using novel local spectral ternary patterns

Automated myocardial infarction identification based on interbeat variability analysis of the photoplethysmographic data

Photoplethysmographic (PPG) measurements from ambulatory subjects may suffer from unreliability due to body movements and missing data segments due to loosening of sensor. This paper describes an on-device reliability assessment from PPG measurements using a stack denoising autoencoder (SDAE) and multilayer perceptron neural network (MLPNN). The missing segments were predicted by a personalized convolutional neural network (CNN) and long-short term memory (LSTM) model using a short history of the same channel data. Forty sets of volunteers’ data, consisting of equal share of healthy and cardiovascular subjects were used for validation and testing. The PPG reliability assessment model (PRAM) achieved over 95% accuracy for correctly identifying acceptable PPG beats out of total 5000 using expert annotated data. Disagreement with experts’ annotation was nearly 3.5%. The missing segment prediction model (MSPM) achieved a root mean square error (RMSE) of 0.22, and mean absolute error (MAE) of 0.11 for 40 missing beats prediction using only four beat history from the same channel PPG. The two models were integrated in a standalone device based on quad-core ARM Cortex-A53, 1.2 GHz, with 1 GB RAM, with 130 MB memory requirement and latency ∼0.35 s per beat prediction with a 30 s frame. The present method also provides improved performance with published works on PPG quality assessment and missing data prediction using two public datasets, CinC and MIMIC-II under PhysioNet.

On-Device Reliability Assessment and Prediction of Missing Photoplethysmographic Data Using Deep Neural Networks

Recently, photoplethysmogram (PPG) signal is widely adopted in health monitoring devices for automated assessment of different cardiovascular parameters. However, research in the area of computeriz...

An Automated Algorithm to Extract Time Plane Features From the PPG Signal and its Derivatives for Personal Health Monitoring Application

Recent advances in cardiovascular diseases (CVDs) have inspired the modern population to adopt different easy-to-use personal health monitoring devices that employ easy-to-acquire biosignal such as Photoplethysmogram (PPG) for regular and long-term monitoring of vital signs. However, the detection of clinically significant PPG fiducial points such as onset and the systolic peak is still an open area of research. In the present work, a robust and simple PPG onset and systolic peak detection algorithm is proposed based on the Hilbert Transform. The methodology uses signal derivative, Hilbert Transform, amplitude thresholding and slope-reversal based approaches. In contrast to other techniques, the proposed algorithm exhibits high efficiency with average sensitivity and positive predictivity of 99.83% and 100% respectively, as tested with PPG records collected from the MIMIC database, institutional laboratory and hospital respectively.

A Robust PPG Onset and Systolic Peak Detection Algorithm Based On Hilbert Transform

Recently, photoplethysmography (PPG)-based techniques have been extensively used for cuff-less, automated estimation of blood pressure because of their inexpensive and effortless acquisition techno...

Ppg-based automated estimation of blood pressure using patient-specific neural network modeling

Life threatening consequences of chronic cardiovascular diseases (CVDs) have motivated the population around the world to opt for regular monitoring of vital signs using smart health monitoring devices. However, such easy-to-use, portable devices are found to be compatible with bio-signals that are easy-to-acquire. Hence, Photoplethysmogram (PPG) signal is popularly adopted in such devices as a competent alternative of existing standard bio-signals. But research in the area of computerized analysis of PPG signal to extract relevant clinical information is still insufficient. In this paper, a robust, automated yet simple feature extraction algorithm is proposed for the PPG signal through accurate detection of characteristic points. The methodology consists of signal derivative, amplitude thresholding slope-reversal and empirical formula based approach. Baseline modulation is removed from the PPG dataset and then features are extracted from the amplitude normalized PPG signal. Performance of the proposed algorithm is evaluated over the standard MIMIC database as well as over real PPG data acquired from both healthy volunteers and cardiac patients. The proposed algorithm offers high efficiency in terms of sensitivity, positive predictivity and low value of errors in the detected features.

Madhuchhanda Mitra

Papers

An Automated Algorithm to Extract Time Plane Features From the PPG Signal and its Derivatives for Personal Health Monitoring Application

Automated myocardial infarction identification based on interbeat variability analysis of the photoplethysmographic data

A Robust PPG Onset and Systolic Peak Detection Algorithm Based On Hilbert Transform

Ppg-based automated estimation of blood pressure using patient-specific neural network modeling

A Robust PPG Time Plane Feature Extraction Algorithm for Health Monitoring Application