Going beyond conventional patient monitoring, homecare systems are being enhanced with affective computing capabilities, which are particularly valuable for both detecting and preventing dangerous situations (for example, intense stress or depression). A patient’s emotional state is known to affect, over time, their health as well as their cognitive abilities, behaviour and quality of life. Through tools and methods for emotion recognition, affective computing aims to create human-centred computing systems that understand the user’s state from audiovisual and biological signals and then interact with the user appropriately. This thesis studies and designs applications for monitoring the emotional and cognitive state of patients. It will examine algorithms for recognising emotions from images (facial photographs) or video, and will use existing smart games, or design new ones, that measure the patient’s mobility and cognitive functions such as memory.
