Enhancing Stress Detection Employing Physiological Signals from the WESAD Dataset: A Machine Learning Approach with SMOTE

2024-12-28T14:00:07+00:00

Numerous medical disorders, including diabetes, heart disease, and hypertension, are significantly influenced by stress. Physiological markers for real-time stress detection are becoming more popular as wearable health monitoring devices are widely used. This research uses the WESAD (Wearable Stress and Affect Detection) dataset, which contains multimodal physiological data, including ECG, EDA, EMG, respiration, and temperature, to predict stress levels. We apply four machine learning classifiers to this dataset, focusing on addressing the class imbalance using the Synthetic Minority Oversampling Technique (SMOTE). The outcomes illustrate that Decision Trees outperform other classifiers with an accuracy of 96.27%. For future work, efforts can be directed towards incorporating additional modalities, such as EEG and eye-tracking data, to improve stress detection accuracy further. Longitudinal data collection could also help understand stress over comprehensive periods, providing insights into chronic stress patterns.

A Detailed Review on Brain Image Segmentation using Deep Learning Algorithms

2024-12-28T14:20:44+00:00

The need for a better method of diagnosis is essential, as evidenced by the increasing incidence of accurate brain tumour identification in the field of neuro-oncology. The existing literature, which is mainly concerned with the classification of MRI images, does not provide thorough answers to the many problems that arise in brain tumour segmentation, including imaging abnormalities, the difficult-to-define boundaries of tumours, tumour heterogeneity, and classification uncertainties. By putting forth a novel deep learning framework that blends the well-known U-Net architecture with self-attention processes, specially tailored for brain tumour segmentation, this study seeks to overcome these problems. Our work thoroughly evaluates and contrasts current deep learning methods, highlighting the efficiency of the U-Net architecture in recognizing both particular and generic patterns in three-dimensional brain imaging. Key findings show that our proposed model outperforms recent advances in brain tumour segmentation from 2020 to 2024 in terms of accuracy, precision, sensitivity, and specificity. Significant results suggest that this combination of factors sets a new standard in medical image segmentation, with the potential to revolutionize diagnostic capabilities and therapeutic approaches. The implications go beyond academic discussion, giving patients and healthcare professionals hope for the accurate diagnosis and management of brain tumours. The integration of self-attention mechanisms has proven effective in improving segmentation accuracy by focusing on critical tumour regions and improving overall precision.

Milestone Transactions on Medical Technometrics

Enhancing Stress Detection Employing Physiological Signals from the WESAD Dataset: A Machine Learning Approach with SMOTE

A Detailed Review on Brain Image Segmentation using Deep Learning Algorithms