Pulmonary Arterial Hypertension Detection from PPG Signals
Problem statement
Cardiovascular diseases are a leading cause of mortality worldwide. Early detection of Pulmonary Arterial Hypertension (PAH) — a severe condition — is critical, as it can improve the 3-to-5-year patient survival rate to over 80–90%. The goal was to build an accessible, non-invasive screening tool using common, low-cost PPG sensors.
My contribution
ML Division Lead
Led the design and implementation of the end-to-end signal processing and classification pipeline as head of the machine learning division.
Signal Pre-processing
Developed pre-processing algorithms to clean raw PPG signals and extract relevant physiological features using Wavelet Scattering Transform.
Model Engineering
Engineered, trained, and validated a Bidirectional LSTM (Bi-LSTM) classification model in Python with TensorFlow/Keras, then ran performance analysis to confirm diagnostic-grade reliability.
Technical stack
Results & impact
- ✓Awarded 1st Place at the West Java Health Tech Innovation Competition 2025.
- ✓Delivered a non-invasive screening approach with potential to push PAH survival rates past 80–90% through earlier detection.
- ✓Built entirely around low-cost, widely available PPG sensors to maximize accessibility.
Have a similar challenge you want to solve?
Start a project discussion →