A5085504440- Emotion and Mood Recognition
- Speech and Audio Processing
- Context-Aware Activity Recognition Systems
- Anomaly Detection Techniques and Applications
- Color perception and design
- Music and Audio Processing
- Electricity Theft Detection Techniques
- Digital Mental Health Interventions
- IoT and Edge/Fog Computing
- Sleep and Wakefulness Research
- Imbalanced Data Classification Techniques
- Face and Expression Recognition
- Sleep and related disorders
Electronics and Telecommunications Research Institute
2018-2022
In this paper, we perform a systematic study about the on-body sensor positioning and data acquisition details for Human Activity Recognition (HAR) systems. We build testbed that consists of eight body-worn Inertial Measurement Units (IMU) sensors an Android mobile device activity collection. develop Long Short-Term Memory (LSTM) network framework to support training deep learning model on human data, which is acquired in both real-world controlled environments. From experiment results,...
To understand the multilateral characteristics of human behavior and physiological markers related to physical, emotional, environmental states, extensive lifelog data collection in a real-world environment is essential. Here, we propose method using multimodal mobile sensing present long-term dataset from 22 subjects 616 days experimental sessions. The contains over 10 000 hours data, including physiological, such as photoplethysmography, electrodermal activity, skin temperature addition...
Speech emotion recognition (SER) is a natural method of recognizing individual emotions in everyday life. To distribute SER models to real-world applications, some key challenges must be overcome, such as the lack datasets tagged with labels and weak generalization model for an unseen target domain. This study proposes multi-path group-loss-based network (MPGLN) support multi-domain adaptation. The proposed includes bidirectional long short-term memory-based temporal feature generator...
Sleep is one of the most important factors in maintaining both physical and mental health. There are many causes sleep problems, it generally necessary to maintain a healthy lifestyle avoid them. In medical field, information related problems including obtained through interviews, but this approach limited because dependent on patient's memory. Thus, there studies adopting ecological momentary assessments (EMAs) collect lifestyles. Some them also use smart devices data effectively. However,...
This paper proposes a SER (Speech Emotion Recognition) framework using user self-referential speech features, which reflect the ROC (Rate of Change) corresponding feature value according to user's emotion states. The experiments shows that proposed features in addition absolute can improve accuracy.
This paper proposes an ensemble classifier based on decision-fusion of multiple SER (Speech Emotion Recognition) models. The one the models used in this work is a typical categorical learning model for classifying emotion labels, while others are A/V (Arousal/Valence) that recognize states Russell's space. evaluation performed shows accuracy proposed combines each output and improved compare to result when applied separately.