StressNAS: Affect State and Stress Detection Using Neural Architecture Search

Huynh, Lam; Nguyen, Tri; Nguyen, Tat Thang; Pirttikangas, Susanna; Siirtola, Pekka

doi:10.1145/3460418.3479320

Cited by 22 publications

(15 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…compared to the original work in[22], employing our proposed bio-signal processing pipeline before feature extraction step increases the mean accuracy score of conventional Machine Learning model (RF model) around 3.2%. Our proposed NN model improves the performance of state-ofthe-art (SOTA) subject-independent stress detection model proposed by Lam et al[13] around 1.63% using the same number of bio-signals. Conventional Machine Learning models also achieve competitive accuracy scores compared to the SOTA model when using the fusion feature of three sensor sources with appropriate signal processing before feature extraction.…”

mentioning

confidence: 70%

“…To improve the accuracy of the subject-independent stress detection model, we conducted experiments on the benchmarking dataset that is used extensively in related works [21,24,13]. The benchmarking dataset named WESAD [22] consists of four different types of low-resolution physiological data collected from 15 participants under two different study protocols in a laboratory environment.…”

Section: Stress Detection Datasetmentioning

confidence: 99%

See 1 more Smart Citation

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

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

mentioning

confidence: 70%

Section: Stress Detection Datasetmentioning

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

View full text Add to dashboard Cite

show abstract

“…To improve the stress prediction accuracy of the subject-independent model, we conducted experiments on the benchmark dataset that is used extensively in related works [22,25,13] to evaluate the performance of stress detection model. The benchmark dataset named WESAD [23] consists of four different types of lowresolution physiological data collected from 15 participants under two different study protocols in a laboratory environment.…”

Section: Stress Detection Datasetmentioning

confidence: 99%

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

Ninh¹,

Nguyen²,

Smyth³

et al. 2022

Preprint

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

“…Gil-Martin et al [30] proposed a CNN-MLP architecture to extract meaningful features from the Fourier transform (FFT), cube root (CR) and constant q spectral transform (CQT) of signal sub-window. Huynh et al [22] used filter bank as model input and automatically selected the optimal model for each modality from 10,000 deep neural networks for training. Finally, features of all modalities were concatenated for classification.…”

Section: Comparison With State-of-the-artmentioning

confidence: 99%

Fusion of Physiological and Behavioural Signals on SPD Manifolds with Application to Stress and Pain Detection

WU,

Daoudi,

Amad

et al. 2022

Preprint

View full text Add to dashboard Cite

Existing multimodal stress/pain recognition approaches generally extract features from different modalities independently and thus ignore cross-modality correlations. This paper proposes a novel geometric framework for multimodal stress/pain detection utilizing Symmetric Positive Definite (SPD) matrices as a representation that incorporates the correlation relationship of physiological and behavioural signals from covariance and cross-covariance. Considering the non-linearity of the Riemannian manifold of SPD matrices, well-known machine learning techniques are not suited to classify these matrices. Therefore, a tangent space mapping method is adopted to map the derived SPD matrix sequences to the vector sequences in the tangent space where the LSTM-based network can be applied for classification. The proposed framework has been evaluated on two public multimodal datasets, achieving both the state-ofthe-art results for stress and pain detection tasks.

show abstract

StressNAS: Affect State and Stress Detection Using Neural Architecture Search

Cited by 22 publications

References 7 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

Fusion of Physiological and Behavioural Signals on SPD Manifolds with Application to Stress and Pain Detection

Contact Info

Product

Resources

About