Detection of a and b waves in the acceleration photoplethysmogram

Table 3 Comparison of different a wave detection performance on the testing dataset (APG signals measured at rest and after 2 hours of exercise)

Algorithm	TP (%)	FN (%)	FP (%)	SE (%)	+P (%)	Threshold values
Algorithm	TP (%)	FN (%)	FP (%)	SE (%)	+P (%)	THR ₁	THR ₂	THR ₃
Proposed	100	0.32	0	99.78	100	M A _beat+α	W ₂	NA
algorithm	100	0.32	0	99.78	100	M A _beat+α	W ₂	NA
AF1	69.5	7.5	30.5	90.25	69.5	0.31	0.0001	-0.001
AF2	0.018	0.27	99.98	6.25	0.018	0.21	0.75	NA
AF3	0	0	100	NaN	0	62	NA	NA
FD1	0.27	2.8	99.73	8.79	0.27	0.099	NA	NA
FD2	0	0	100	NaN	0	150	NA	NA
DF1	0	0	100	NaN	0	21	NA	NA
DF2	48.8	14.2	51.2	77.46	48.8	1	0.06	NA
FS1	2.42	0.3	97.58	88.97	2.42	154.5	NA	NA
FS2	42.46	6.9	57.54	86.02	42.46	0.55	0.47	NA

The PPG signals were collected from 27 subjects for 20 seconds during the 5 minutes break between each exercise [5]. To compare the performance of the proposed algorithm with the nine algorithms [5], two statistical measures were used: SE=TP/(TP+FN) and +P=TP/(TP+FP), where TP is the number of true positives (a wave detected as a wave), FN is the number of false negatives (a wave has not been detected), and FP is the number of false positives (non-a wave detected as a wave). Here, NA stands for Not Applied, while NaN stands for Not-a-Number.

ISSN: 1475-925X