Analysing the Performance of Stress Detection Models on Consumer-Grade Wearable Devices

Ninh, Van-Tu; Smyth, Sinéad; Tran, Minh–Triet; Gurrin, Cathal

doi:10.3233/faia210050

Cited by 10 publications

(7 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The total duration of the study protocol is about two hours, which is considered to be long enough to capture sufficient physiological data for stress detection model training. Since previous works using this dataset employ study protocol is used as the ground-truth for constructing the stress detection model [22,23,25,21], we also use the same ground-truth as in previous works for consistent comparison of stress prediction accuracy of the proposed models. In detail, the baseline and amusement condition are classified into non-stress class while the stress condition is considered as the stress one.…”

Section: Stress Conditionmentioning

confidence: 99%

“…For the BVP, we firstly clean the raw signal in each window segment by removing the outlier values over the 98 th and below the 2 th percentile using winsorisation method as in [5] and removing the baseline drift using Butterworth high-pass filter with cut-off frequency of 0.5 Hz as in [14]. We then apply min-max normalization to the cleaned signal to minimise the physiological signal difference between individuals before following the previous research [21] to employ the Elgandi processing pipleline [4] to clean the photoplethysmogram (PPG) signal [19] and detect systolic peaks. The systolic peaks are used to compute a list of RR-intervals, which are then pre-processed using the hrv-analysis package to remove outliers and ectopic beats [28] as well as interpolating missing values.…”

Section: Bio-signal Processing and Statistical Feature Extraction Of ...mentioning

confidence: 99%

See 1 more Smart Citation

An Improved Subject-Independent Stress Detection Model Applied to Consumer-grade Wearable Devices

Ninh¹,

Nguyen²,

Smyth³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Stress is a complex issue with wide ranging physical and psychological impacts on human daily performance. Specifically, acute stress detection is becoming a valuable application in contextual human understanding. Two common approaches to training a stress detection model are subject-dependent and subject-independent training methods. Although subject-dependent training methods have proven to be the most accurate approach to build stress detection models, subject-independent models are a more practical and cost-efficient method, as they allow for the deployment of stress level detection and management systems in consumer-grade wearable devices without requiring training data for the end user. To improve the performance of subject-independent stress detection models, in this paper, we introduce a stress-related bio-signal processing pipeline with a simple neural network architecture using statistical features extracted from multimodal contextual sensing sources including Electrodermal Activity (EDA), Blood Volume Pulse (BVP), and Skin Temperature (ST) captured from a consumer-grade wearable device. Using our proposed model architecture, we compare the accuracy between stress detection models that use measures from each individual signal source, and one model employing the fusion of multiple sensor sources. Extensive experiments on the publicly available WESAD dataset demonstrate that our proposed model outperforms conventional methods as well as providing 1.63% higher mean accuracy score compared to the state-of-the-art model while maintaining a low standard deviation. Our experiments also show that combining features from multiple sources produces more accurate predictions than using only one sensor source individually.

show abstract

Section: Stress Conditionmentioning

confidence: 99%

Section: Bio-signal Processing and Statistical Feature Extraction Of ...mentioning

confidence: 99%

An Improved Subject-Independent Stress Detection Model Applied to Consumer-grade Wearable Devices

Ninh¹,

Nguyen²,

Smyth³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…The total duration of the study protocol is about two hours, which is considered to be long enough to capture sufficient physiological data to train a stress detection model. Since previous works on this dataset employ study-protocol as the ground-truth of both train and test data [21,22,24,20], we also use the same ground-truth construction method as in previous works for consistent comparison of the results. In detail, the baseline and amusement condition are classified into non-stress class while the stress condition is considered as the stress one.…”

Section: Stress Conditionmentioning

confidence: 99%

“…For the BVP, we firstly clean the raw signal in each window segment by removing the outlier values over the 98 th and below the 2 th percentile using winsorisation method as in [5] and removing the baseline drift using Butterworth high-pass filter with cut-off frequency of 0.5 Hz as in [14]. We then apply min-max normalization to the cleaned signal to minimise the physiological signal difference between individuals before following the previous research [20] to employ the Elgandi processing pipleline [4] for the photoplethysmogram (PPG) signal clearning [18] and the systolic peaks detection. The systolic peaks are used to compute a list of RR-intervals, which are then pre-processed using the hrvanalysis package to remove outliers and ectopic beats [27] as well as interpolating missing values.…”

Section: Bio-signal Processing and Statistical Feature Extraction Of ...mentioning

confidence: 99%

An Improved Subject-Independent Stress Detection Model Applied to Consumer-grade Wearable Devices

Ninh

Nguyen

Smyth

et al. 2022

Advances and Trends in Artificial Intelligence. Theory and Practices in Artificial Intelligence

Self Cite

View full text Add to dashboard Cite

Stress is a complex issue with wide-ranging physical and psychological impacts on human daily performance. Specifically, acute stress detection is becoming a valuable application in contextual human understanding. Two common approaches to training a stress detection model are subject-dependent and subject-independent training methods. Although the subject-dependent training method is proven to be the most accurate approach to build stress detection models, subjectindependent one is a more practical and cost-efficient method, as it facilitates the deployment of stress level detection and management systems in consumer-grade wearable devices without requiring additional data for training from end-users. To improve the performance of subjectindependent stress detection models, in this paper, we introduce a stressrelated bio-signal processing pipeline with a simple neural network architecture using statistical features extracted from multimodal contextual sensing sources including Electrodermal Activity (EDA), Blood Volume Pulse (BVP), and Skin Temperature (ST) captured from a consumergrade wearable device. Using our proposed model architecture, we compare the accuracy of stress detection models that use measures from each individual signal source with the one employing the fusion of multiple sensor sources. Extensive experiments on the publicly available WESAD dataset demonstrate that our proposed model outperforms conventional methods as well as providing 1.63% higher mean accuracy score compared to the state-of-the-art model while maintaining a low standard deviation. Our experiments also show that combining features from multiple sources produces more accurate predictions than using only one sensor source individually.

show abstract

“…The HR and HRV measurements captured by the sensors not only provide information about physical health, they can also help track mental stress which has secondary deleterious effects on health, including mental health. 4 – 7 The most commonly used sensors for cardiovascular measurements are wrist-worn smart watches; however, chest strap sensors are also widely used, especially in the context of sports. More recently, a class of devices that are designed to be worn on the upper arm or the forearm have become commercially available.…”

Section: Introductionmentioning

confidence: 99%

Feasibility of Using an Armband Optical Heart Rate Sensor in Naturalistic Environment

Kotlyar

Dufresne

et al. 2022

Preprint

View full text Add to dashboard Cite

Consumer-grade heart rate (HR) sensors including chest straps, wrist-worn watches and rings have become very popular in recent years for tracking individual physiological state, training for sports and even measuring stress levels and emotional changes. While the majority of these consumer sensors are not medical devices, they can still offer insights for consumers and researchers if used correctly taking into account their limitations. Multiple previous studies have been done using a large variety of consumer sensors including Polar devices, Apple watches, and Fitbit wrist bands. The vast majority of prior studies have been done in laboratory settings where collecting data is relatively straightforward. However, using consumer sensors in naturalistic settings that present significant challenges, including noise artefacts and missing data, has not been as extensively investigated. Additionally, the majority of prior studies focused on wrist-worn optical HR sensors. Arm-worn sensors have not been extensively investigated either. In the present study, we validate HR measurements obtained with an arm-worn optical sensor (Polar OH1) against those obtained with a chest-strap electrical sensor (Polar H10) from 16 participants over a 2-week study period in naturalistic settings. We also investigated the impact of physical activity measured with 3-D accelerometers embedded in the H10 chest strap and OH1 armband sensors on the agreement between the two sensors. Overall, we find that the arm-worn optical Polar OH1 sensor provides a good estimate of HR (Pearson r = 0.90, p < 0.01). Filtering the signal that corresponds to physical activity further improves the HR estimates but only slightly (Pearson r = 0.91, p < 0.01). Based on these preliminary findings, we conclude that the arm-worn Polar OH1 sensor provides usable HR measurements in daily living conditions, with some caveats discussed in the paper.

show abstract

Analysing the Performance of Stress Detection Models on Consumer-Grade Wearable Devices

Cited by 10 publications

References 21 publications

An Improved Subject-Independent Stress Detection Model Applied to Consumer-grade Wearable Devices

An Improved Subject-Independent Stress Detection Model Applied to Consumer-grade Wearable Devices

An Improved Subject-Independent Stress Detection Model Applied to Consumer-grade Wearable Devices

Feasibility of Using an Armband Optical Heart Rate Sensor in Naturalistic Environment

Contact Info

Product

Resources

About