Skip to main content
. 2021 Sep 8;21(18):6022. doi: 10.3390/s21186022

Figure A4.

Figure A4

Left: Mean absolute error (MAE) of subjects that did not show any improvement before vs. after fine tuning using remote photoplethymographic (rPPG) data for most of the architectures tested. Right: Comparison of the MAE for different personalization strategies using the first 20% of the test subject’s data vs. randomly drawn data for AlexNet (a) and ResNet (b).