Payal Mohapatra

Bio: Payal Mohapatra is an academic researcher from Indian Institute of Technology Madras. The author has contributed to research in topics: Artificial intelligence & Speech recognition. The author has an hindex of 2, co-authored 2 publications receiving 20 citations.

A novel sensor for wrist based optical heart rate monitor

Abstract: The growing demands of continuous healthcare and hence physiological monitoring necessitates a system with high reliability and accuracy. Wearable used for continuous cardiological parameter estimation from wrist use reflective photoplethysmography technique that has certain limitations which are imperative. One such constraint is skin pigmentation of the subject. In the present work a sensor module design is proposed addressing to the anomalies due to optical properties of skin. A novel 590 nm (yellow-orange) wavelength based optical system is tailored suitably to maximize the signal quality acquired. The proposed setup is validated on a conglomeration of subjects in terms of age, gender and skin tone. A generous agreement between coherent measures for signal quality shows that the proposed wavelength holds an advantage over its shorter counterpart when subjected to varied skin pigmentation levels. A maximum improvement factor of 71 is observed in case of perfusion index, 31 for pulsatile strength and 3 for SNR. The details of sensor design, experimental setup, validation protocol, observations and inferences drawn from the study are presented.

A Yellow–Orange Wavelength-Based Short-Term Heart Rate Variability Measurement Scheme for Wrist-Based Wearables

Abstract: Heart rate variability (HRV) is one of the important biomarkers of physical and psychological well-being. Hence, a convenient and minimally intrusive method for HRV measurement is advantageous. Although high levels of surrogacy of short-term HRV estimates obtained from the measurements of blood volume changes to traditional electrocardiographic (ECG) measurements have been reported, no detailed account on extraction of such parameters from a wrist-based optical monitor is found in the literature. In this paper, a yellow–orange wavelength-based optical scheme is incorporated into a wearable device for HRV estimation from dorsal side of the wrist. This design is pivotal in catering to a wider span of population with varied skin tones. The developed wearable in alliance with a gateway device is capable of picking up photoplethysmography from the measurement site, allowing estimation of HRV-indices within a confidence of 5% from ECG-derived parameters. The HRV measurement ecosystem is validated under the setting of three postural loads for 20 subjects, generating 60 data sets. Study results show statistically significant positive correlation and nonsignificant bias in Bland–Altman analysis, for the HRV-indices derived from either method. In most of the extracted HRV features, the observations in supine position showed minimum deviation from the reference. Estimation of short-term HRV-indices from wrist-based photoplethysmography under stationary conditions shows promising results from the study. Electrical and biological noninterference and ease of usage of the proposed design simplify stationary and ambulatory monitoring of HRV.

Efficient Stuttering Event Detection Using Siamese Networks

Abstract: Speech disfluency research is pivotal to accommodating atypical speakers in mainstream conversational technology. However, the lack of publicly available labeled and unlabeled datasets is a significant bottleneck to such research. While many works use pseudo dysfluency data with proxy labels and formulate a self-supervised task, we see merit in using real-world data. In this work, we consolidate the corpora of publicly available speech disfluency datasets with and without labels and propose DisfluentSiam – an efficient siamese network-based small-scale pretraining pipeline using task-specific data from multiple domains with only 10M trainable parameters. We show that with DisfluentSiam, we achieve an average of 15% boost in performance across five types of dysfluency event detection compared to direct wav2vec 2.0 embeddings. In particular, with only 4-5 mins of labeled data for fine-tuning, the DisfluentSiam demonstrates the advantage of task-specific pretraining with up to 25% higher accuracy.

Person Identification with Wearable Sensing Using Missing Feature Encoding and Multi-Stage Modality Fusion

Abstract: We present a missingness-aware fusion network (MAFN) to identify a person’s digital phenotype from continuously measured longitudinal multi-modal wearable data. This work is done as a part of Track 1 of e-Prevention: Person Identification and Relapse Detection from Continuous Recordings of Biosignals Signal Processing Grand Challenge at International Conference on Acoustics, Speech, & Signal Processing (ICASSP) 2023. MAFN achieves an accuracy of 91.36% on test data. Additionally, our experiments confirm findings from previous works that kinetic features derived from the accelerometer in-deed contain more discriminative features for person identification task.

Sensors and Functionalities of Non-Invasive Wrist-Wearable Devices: A Review

Abstract: Wearable devices have recently received considerable interest due to their great promise for a plethora of applications. Increased research efforts are oriented towards a non-invasive monitoring of human health as well as activity parameters. A wide range of wearable sensors are being developed for real-time non-invasive monitoring. This paper provides a comprehensive review of sensors used in wrist-wearable devices, methods used for the visualization of parameters measured as well as methods used for intelligent analysis of data obtained from wrist-wearable devices. In line with this, the main features of commercial wrist-wearable devices are presented. As a result of this review, a taxonomy of sensors, functionalities, and methods used in non-invasive wrist-wearable devices was assembled.

Pulse rate variability in cardiovascular health: a review on its applications and relationship with heart rate variability

Abstract: Heart rate variability has been largely used for the assessment of cardiac autonomic activity, due to the direct relationship between cardiac rhythm and the activity of the sympathetic and parasympathetic nervous system. In recent years, another technique, pulse rate variability, has been used for assessing heart rate variability information from pulse wave signals, especially from photoplethysmography, a non-invasive, non-intrusive, optical technique that measures the blood volume in tissue. The relationship, however, between pulse rate variability and heart rate variability is not entirely understood, and the effects of cardiovascular changes in pulse rate variability have not been thoroughly elucidated. In this review, a comprehensive summary of the applications in which pulse rate variability has been used, with a special focus on cardiovascular health, and of the studies that have compared heart rate variability and pulse rate variability is presented. It was found that the relationship between heart rate variability and pulse rate variability is not entirely understood yet, and that pulse rate variability might be influenced not only due to technical aspects but also by physiological factors that might affect the measurements obtained from pulse-to-pulse time series extracted from pulse waves. Hence, pulse rate variability must not be considered as a valid surrogate of heart rate variability in all scenarios, and care must be taken when using pulse rate variability instead of heart rate variability. Specifically, the way pulse rate variability is affected by cardiovascular changes does not necessarily reflect the same information as heart rate variability, and might contain further valuable information. More research regarding the relationship between cardiovascular changes and pulse rate variability should be performed to evaluate if pulse rate variability might be useful for the assessment of not only cardiac autonomic activity but also for the analysis of mechanical and vascular autonomic responses to these changes.

Evaluation of the signal quality of wrist-based photoplethysmography

Abstract: Objective Wearable devices with embedded photoplethysmography (PPG) sensors enable continuous monitoring of cardiovascular activity, allowing for the detection cardiovascular problems, such as arrhythmias. However, the quality of wrist-based PPG is highly variable, and is subject to artifacts from motion and other interferences. The goal of this paper is to evaluate the signal quality obtained from wrist-based PPG when used in an ambulatory setting. Approach Ambulatory data were collected over a 24 h period for 10 elderly, and 16 non-elderly participants. Visual assessment is used as the gold standard for PPG signal quality, with inter-rater agreement evaluated using Fleiss' Kappa. With this gold standard, 5 classifiers were evaluated using a modified 13-fold cross-validation approach. Main results A Random Forest quality classification algorithm showed the best performance, with an accuracy of 74.5%, and was then used to evaluate 24 h long ambulatory wrist-based PPG measurements. Significance In general, data quality was high at night, and low during the day. Our results suggest wrist-based PPG may be best for continuous cardiovascular monitoring applications during the night, but less useful during the day unless methods can be identified to improve low quality signal segments.