Overview
Heart sounds are a first check for cardiovascular disease, but conventional stethoscopes depend heavily on the listener’s experience and cannot record or share what they hear. This project designs a 3D-printed digital stethoscope built around MEMS sensors, so heart sounds can be captured, analysed and shared in real time.
What we built
- MEMS stethoscope. A 3D-printed body with MEMS sensors that capture a broad range of heart sound frequencies, and an embedded board for signal conditioning and denoising.
- ML model. Heart sounds are classified as normal or abnormal. Models were trained on public phonocardiogram datasets and then on recordings from our own prototype.
- Mobile app and server. The app streams the heart sound as a time-frequency waveform and produces an automated report. A secure backend stores the data.
- Telemedicine. A remote doctor can follow the live signal alongside the ML report.
Goal
Earlier screening and faster diagnosis for cardiovascular disease, using affordable and more sustainable materials, with the prototype tested on volunteers under ethical approval.