Motion-artifact mitigation
Developing methods for denoising PPG and ECG signals while preserving physiologically meaningful information, including beat timing, HRV, and waveform morphology.
Research
Our research addresses four interconnected areas in robust physiological signal analysis, with the goal of improving reliable inference from real-world biosignals.
Signal quality assessment
Developing signal-quality assessment methods to identify unreliable segments and improve the robustness of downstream physiological inference.
Wearable cardiovascular monitoring
Investigating physiological parameter estimation and cardiovascular monitoring from wearable and smartphone-based sensing, including heart-rate estimation, arrhythmia classification, and non-invasive glucose assessment.
Transformers & physiological modeling
Applying attention-based architectures and physiological modeling to capture complex temporal dynamics, improve robustness to motion and measurement variability, and support controlled evaluation of learning-based methods.