This respository hosts the research for the manuscript entitled, "Data-Driven Guided Attention for Analysis of Physiological Waveforms with Deep Learning". The repo is organized as follows:
- Data Annotation - Hosts the files necessary for annotating the fiducial points for both the Bio-Z and MIMIC PPG modalities with Boosted-SpringDTW.
- DDGA - Hosts the files necessary for running DDGA for the Personalized experiments and the Interpolation and Extrapolation experiments.
- DDGA/Personalized/ - These experiments train on a given subjects data and test on the same subjects data.
- DDGA/Interploation and Extrapolation/ - These experiments train on specific ranges of a person's BP data and tests on the remaining BP ranges.
DDGA/Personalized/ and DDGA/Interploation and Extrapolation/ contain the DDGA codes for each evaluated DL type for predicting DBP/SBP separately, for both Bio-Z and MIMIC.