This project addresses to estimate heart rate (HR) during exercise in real-time using wrist-type PPG signals amidst intense motion artifacts. We used multi-channel adpative filtering to denoise the corrupted-ppgs, then found the HR peaks using periodogram and HR tracking.
This problem was the IEEE SP CUP 2015 challenge problem and our algo became 1st among all the teams and was the state-of-the-art at that time.
The work was published in IEEE JBHI- https://ieeexplore.ieee.org/abstract/document/7755741
The codes attached here were implemented in MATLAB, let us know your comments/suggestions or if you find any bugs at chowdh23@purdue.edu