Facial movement dysfunctions: Conceptual design of a therapy-accompanying training system

Abstract:
In this work, we present the scenario of a camera-based training system for patients with dysfunctions of facial muscles. The system is to be deployed accompanying to therapy in a home environment. The aim of the intended application is to support the unsupervised training sessions and to provide feedback. Based on conversations with speech-language therapists and the analysis of existing solutions, we derived a theoretic model that facilitates the conceptual design of such an application. Furthermore, the work is concerned with implementation details, with main focus on the automatisation of the face analysis. We motivate the selection of the features and examine their discriminative power and robustness for the automated recognition of therapeutic facial expressions in a real-world application.