The DiaVoc project: Diagnosing vocal characteristics to track patients’ health

Andreas Jakobsson, LU & Johan Sanmartin Berglund, BTH

This project focuses on the diagnosis and tracking of health conditions that affect patients’ voices, such as Neurocognitive disorders (NCDs) (cognitive decline), pulmonary disorder (COPD), and heart failure conditions (HF). Using longitudinal voice recordings, matched with medical information, we will develop mathematical vocal features, distance measures, and machine learning techniques suitable for tracking and classifying relevant vocal changes. These features will improve clinical assessments of underlying health conditions and increase treatment efficacy and prediction of prognosis. A key strength of the project are our transdisciplinary perspectives in the cross-section between health science and engineering technology.

Project number: C07