From affordable personal informatics technologies such as smartphones and tablet computers, new possibilities to support treatment of patients with chronic diseases have arisen. In particular, smartphones enable recording and monitoring of symptoms in near real-time and therefore provide a low-cost and scalable scope for a sensing application, which is able to trigger evidence based health interventions for people with chronic diseases.
In this work, we focus on development of a machine learning model for speaker verification from smartphone audio recordings of asthmatic symptoms, such as cough. Speaker verification would enable an unambiguous assignment of symptoms to one patient.
The successful applicant will first conduct a literature review and subsequently identify promising approaches to solving the problem. Subsequently, he will implement, optimize and evaluate these methods on data collected in a lab study. Ideally, in a future step, this model will be deployed in form of an Android mobile application.
From affordable personal informatics technologies such as smartphones and tablet computers, new possibilities to support treatment of patients with chronic diseases have arisen. In particular, smartphones enable recording and monitoring of symptoms in near real-time and therefore provide a low-cost and scalable scope for a sensing application, which is able to trigger evidence based health interventions for people with chronic diseases. In this work, we focus on development of a machine learning model for speaker verification from smartphone audio recordings of asthmatic symptoms, such as cough. Speaker verification would enable an unambiguous assignment of symptoms to one patient.
The successful applicant will first conduct a literature review and subsequently identify promising approaches to solving the problem. Subsequently, he will implement, optimize and evaluate these methods on data collected in a lab study. Ideally, in a future step, this model will be deployed in form of an Android mobile application.
In the current project we focus on the development of a cough detection model for the smartphone. This works builds upon previous works and focus on building a speaker verification model in Python.
For the application, you should meet the following requirements: Experience in programming Skills or the interest to learn the basics of deep/ machine learning Interest in acoustics and mobile technology ETH / UZH students are preferred
In the current project we focus on the development of a cough detection model for the smartphone. This works builds upon previous works and focus on building a speaker verification model in Python.
For the application, you should meet the following requirements: Experience in programming Skills or the interest to learn the basics of deep/ machine learning Interest in acoustics and mobile technology ETH / UZH students are preferred
Contact Details
Please send your application with a short motivation letter (~250 words), transcript of records and CV to Filipe Barata (fbarata@ethz.ch).
Contact Details Please send your application with a short motivation letter (~250 words), transcript of records and CV to Filipe Barata (fbarata@ethz.ch).