In the evolving field of Speech Emotion Recognition (SER), essential for understanding and addressing mental health issues, conventional models often falter in interpreting complex emotional states, particularly those...
详细信息
In the evolving field of Speech Emotion Recognition (SER), essential for understanding and addressing mental health issues, conventional models often falter in interpreting complex emotional states, particularly those related to mental health conditions like ptsd. This study introduces the Cognitive Emotion Fusion Network (CEFNet), a novel hybrid SER model integrating Improved and Faster Region-based Convolutional Neural Networks (IFR-CNN), Deep Convolutional Neural Networks (DCNNs), Deep Belief Networks (DBNs), and the Bird ' s Nest Learning Analogy (BNLA). Aimed at surpassing the limitations of traditional models, CEFNet focuses on accurately interpreting nuanced emotional expressions, employing advanced machine learning techniques and comprehensive feature extraction. Evaluated using the EMODB and RAVDESS datasets, CEFNet demonstrated superior performance, achieving an accuracy of 98.11% and 91.17% on these datasets, respectively, outperforming existing models in precision and F1 scores. This research marks a significant contribution to SER, particularly in mental health applications, offering a robust framework for emotion recognition in speech. It opens avenues for future enhancements, including broader applicability across languages and cultural contexts, optimization for resource-limited environments, and integration with other modalities for more holistic emotion recognition.
Post-traumatic stress disorder is a major public health concern with a lifetime prevalence rate of 6.1-9.2% in North America. ptsd is known to alter the autonomic nervous system leading to chronic sympathetic arousal ...
详细信息
ISBN:
(纸本)9781728176055
Post-traumatic stress disorder is a major public health concern with a lifetime prevalence rate of 6.1-9.2% in North America. ptsd is known to alter the autonomic nervous system leading to chronic sympathetic arousal including heightened anxiety and hypervigilance. Pupillometry offers a quick and accessible measure of autonomic nervous system imbalances characteristic of ptsd. This study investigates the utility of pupillometry as a biomarker to detect ptsd in a sample of 39 adults with (n = 22) and without (n = 17) ptsd. Participants viewed a 25-minute computer protocol consisting of 5-minute rest phase, 10-minute negative emotionally valent images, and 10-minute guided meditation. We relied on a time-frequency analysis to represent the pupillary responses of two different groups (ptsd-affected individuals and healthy-control subjects). These data were then employed with a CNN network to learn a prediction model. Individuals with ptsd demonstrated increased pupil dilation across the entire protocol. The final outcome revealed an accuracy of 81.09% which indicates the feasibility of using this approach to detecting participants with ptsd in an automated way. Findings from this research have important implications for clinical mental health assessment, diagnostics and treatment.
Post-traumatic stress disorder (ptsd) can be a debilitating condition and early intervention can be instrumental in preventing patients' suffering. Identifying patients at risk for ptsd is challenging because of t...
详细信息
ISBN:
(纸本)9781450376990
Post-traumatic stress disorder (ptsd) can be a debilitating condition and early intervention can be instrumental in preventing patients' suffering. Identifying patients at risk for ptsd is challenging because of the limitations of the available data set, variations in the symptoms of ptsd for different patients, and misdiagnosis due to symptoms being shared with other conditions. In this preliminary study, we explore a small set of structured primary care data extracted from patients' electronic medical records (EMR) from Manitoba, Canada. This data has a small subset of ptsd positive cases, and is used to assess the feasibility of applying machine learning algorithms to diagnose ptsd. We developed three supervised machine learning models, a multilayered perceptron artificial neural network (ANN) model, a support vector machine (SVM), and a random forest classifier (RF) to identify ptsd patients using 890 patients' records. These methods obtained 0.79, 0.78, and 0.83 AUC respectively, which are better than all of the previous work that used EMR data having comparable size as our data. This study is geared towards understanding the primary care standard for ptsd patients in Canada in general and military-veteran population and developing a case definition for ptsd. This initial result demonstrates that an automated ptsd screening tool can be developed based on historical medical data for further study. In our ongoing work, we are exploring the providers' chart notes from the EMR data, which is unstructured text data, to improve the model accuracy and understand the progression of ptsd.
暂无评论