Deep Learning versus Professional Healthcare Equipment: A Fine-Grained Breathing Rate Monitoring Model

Table 1

The results from 6 models: support vector machine (SVM), logistic regression (LR), DRNN (baseline), Model 1, Model 2, and Model 3. All of them are trained and tested on two data sets. The rightmost 4 columns are the average scores over the 4 distances.

	20 cm (0.8 m)		40 cm (0.95 m)		60 cm (1.1 m)		80 cm (1.25 m)
	TPR TNR		TPR TNR		TPR TNR		TPR TNR		TPR	TNR	WAR	TTR (bmp)
SVM	0.897	0.923	0.787	0.920	0.402	0.843	0.316	0.820	0.601	0.876	0.807	0.53
LR	0.867	0.913	0.757	0.932	0.413	0.789	0.342	0.829	0.594	0.865	0.798	0.53
Baseline	0.994	0.988	0.948	0.916	0.509	0.925	0.461	0.947	0.715	0.944	0.886	0.64
DeepFilter 1	0.996	0.970	0.980	0.986	0.667	0.932	0.639	0.921	0.820	0.952	0.919	0.71
DeepFilter 2	0.998	0.965	0.981	0.988	0.740	0.910	0.703	0.900	0.855	0.943	0.921	0.77
DeepFilter 3	0.998	0.976	0.998	0.976	0.733	0.900	0.710	0.921	0.859	0.943	0.922	0.77