Master's Thesis

Development of a Speech Recognition System for Teaching Clinical Skills in Diagnosing Mental Disorders


Yustinus Aquila Adrian


Processing Period:

09.01.2023 - 10.07.2023


In the education of psychiatrists, actors or actresses are often used as simulation patients to replicate clinical situations. However, this method is often limited by resources and time. Therefore, Virtual Reality (VR) can provide a possible solution. The state of the application used for this thesis includes a conversation between a virtual psychiatrist and a virtual patient. A user can only sit inside the virtual environment and listen to the conversation. However, no interaction can be done inside the application. This thesis aims to incorporate interactivity into the conversation between a player as a virtual psychiatrist and a virtual patient. A crucial feature that the current state-of-the-art has not been able to facilitate. This master thesis proposes an application extension to complement existing psychiatrist communication training using speech recognition. There is also a progress tracker to track players’ progress and a multilevel hint system to help players in case they are stuck during training. The proposed implementation can recognize German native speakers’ utterances and intentions. Field evaluations showed this voice recognition works well, with a stable internet connection being the sole requirement. A beginner user with little or no prior experience diagnosing mental illnesses will benefit immensely from a multilevel hint system and a progress tracker to train their communication skills.