Virtual Human Technologies for Multimodal Mental Health Assessment
This project explores the use of virtual humans, also known as Embodied conversation agents (ECAs), to facilitate machine-assisted assessment of depression and anxiety. With advancements in Affective Computing, ECAs now have the ability to recognise and generate human-like behaviours and expressions. It is now also possible to objectively track behavioural and biomedical biomarkers of depression and anxiety. This project leverages both capabilities to simulate naturalistic human interactions where an ECA engages a user to complete computer-based tasks while tracking their natural responses and activity. The project has two main objectives. First, it investigates the effectiveness and impact of using virtual human technologies in automated mental health assessment. Secondly, it aims to develop novel multimodal deep-learning techniques for predicting the severity of self-reported anxiety and depression. Preliminary findings can be found here: Design and Evaluation of Virtual Human Mediated Tasks for Assessment of Depression and Anxiety.
CVL people: Joy Egede
