Performance Analysis of Ten Common QRS Detectors on Different ECG Application Cases

Table 1

The list of six databases.

Database	Description	Number of beats	Number of records	Record length (min)	Total time (min)	Sample frequency (Hz)	Source
A	High-quality ECGs	72,415	100	10	1000	250	2014 PhysioNet/CinC challenge training set (https://physionet.org/challenge/2014/)
	Low-quality ECGs	78,618	100	10	1000	360	2014 PhysioNet/CinC challenge augmented training set (https://physionet.org/challenge/2014/)
B	Normal subjects	1,806,792	18	120	2160	500	MIT-BIH NSR database (https://physionet.org/physiobank/database/nsrdb/)
	Arrhythmia patients	103,724	44	30	1320	360	MIT-BIH arrhythmia database (https://www.physionet.org/physiobank/database/mitdb/)
C	Paced rhythm ECGs	8923	4	30	120	360	MIT-BIH arrhythmia database (https://www.physionet.org/physiobank/database/mitdb/)
D	Telehealth environment ECGs	6708	250	0.5	125	500	Harvard dataverse TELE database (https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/QTG0EP)
Total	—	2,077,180	516	—	5725	—	—