PulseAI - our machine learning algorithm for fiducial point detection - is now published in BSPC.

Excited to share the our work on PulseAI, a machine learning algorithm for fiducial point detection in cardiac waveforms, has been pusblished in [Biomedical Signal Processing and Control].

With the rising prevalence of pulse waveform analysis in both academic and clinical practice, it is of critical importance to correctly identify key timing events within the cardiac waveform. These fiducial points serve as the precursors for measurement of various cardiac indices on the pressure waveform, such as arterial stiffness. Therefore their correct and repeatable identification is important for widespread implementation. In this study, we trained a machine learning model to locate fiducial points, both the inflection point and dicrotic notch, on the cardiac waveform captured at the brachial site. We showed that our machine learning approach successfully identifies these fiducial points within acceptance criteria and can be used to accurately measure parameters such as Augmentation Index automatically.