Consider human activity recognition-related intelligent sensor systems. Several human-centered applications, including active and assisted living (AAL), video surveillance systems, various forms of human–computer interaction, and robotics for human behavior characterization, require a reliable activity recognition system. However, recognizing human activities, either from visual sensors (e.g., video sequences or still images) or from non-visual sensors (e.g., sound sensors, movement detection sensors, etc.), is a relatively challenging task due to a series of problems, such as background clutter, partial occlusion, changes in scale, viewpoint, lighting, appearance, etc.
Reliable intelligent sensor systems are crucial for AAL. Indeed, the field of active and assisted living has evolved dramatically in the last few decades. AAL focuses on providing support to people, especially old or disabled persons, primarily in their smart environments to make their life easier. However, this support is not only limited to patients or old persons, but can also include their relatives, social support services, health workers, and care givers.
Not only in the context of AAL, but also for various other application contexts, emotion and/or stress recognition sensor systems are important system components. Different technologies are used to collect emotion-relevant data. Examples include: physiological sensors (e.g., EEG, ECG, electrodermal activity and skin conductance), and various non-intrusive sensors (e.g., piezo-vibration sensors, facial images, chairborne differential vibration sensors, bed-borne or chair-borne differential vibration sensors). Selected application contexts for sensors and stress-related sensor systems range from driver assistance systems, medical patient monitoring systems, and diverse emotion-aware intelligent systems, up to complex collaborative robotics systems.
Selected keywords (not limited to):
Context modeling for active and assisted living (AAL)-related sensor systems: ontologies, conceptual models, cognitive models, etc.
Human actions and complex activity recognition involving various sensors: visual sensors, non-visual sensors, multimodal sensors, etc.
Human emotion recognition involving various sensor types: visual sensors, physiological sensors, multimodal sensors, etc.
Human support for AAL-related sensor systems
Machine learning and neurocomputing techniques for sensor systems to robustly recognize and predict both emotion and stress: graphical models, neural network methods (e.g., LSTM networks, cellular neural networks, RNN, etc.), deep learning methods, statistical learning, multivariate empirical mode decomposition, etc.
Sensor system concepts for robust subject-independent emotion and stress recognition
Prof. Dr. Kyandoghere Kyamakya
Dr. Fadi Al-Machot
Dr. Ahmad Haj Mosa
Dr. Jean Chamberlain Chedjou
Prof. Dr. Zhong Li
Prof. Dr. Antoine Bagula