Table 3.

A summary of the related works.

Author	Segmentation	Feature	Classification	Data Source	Scope Size	Classification Type	Enrollment to Classification Time	Enrollment Length	Classification Length
[1]	No need	Fiducial	SIMCA	Private	20	Identification	Not specified	Not specified	Not specified
[11]	No need	DWT	Correlation coefficient	Public	18–48	Identification	Not specified	Not specified	Not specified
[8]	Heartbeat	Fiducial	Similarity thresholding	Public	73	Verification	Not specified	30 s	4 s
[14]	R-R and R-peak with fixed length	Fiducial and DWT	Euclidean distance and DTW	Public	20–1119	Verification	Not specified	10 segments	Vary depends on database
[19]	Heartbeat	Fiducial	PCA and Euclidean distance	Public	100	Verification	Not specified	30 s	30 s
[5]	R-peak with fixed length	Fiducial and DWT	Random forest and wavelet distance	Public	18–89	Identification	Not specified	67% of extracted heartbeats	1 heartbeat
[21]	R-peak with fixed length	Learned	CNN	Public	18–23	Identification	Not specified	100 heartbeats	3 heartbeats
[6]	Blind with fixed length	Multi-domain, MFCC, FFT, DCT, Teager, etc.	Channel attention module (CNN)	Public	50–89	Identification	Avg. 63 days for PTBDB, avg. 9 days for EDGIDDB	40 s	4 s
[2]	Heartbeat	DWT	NEWFM	Public	73	Verification	Not specified	15 heartbeats	1 heartbeat
[3]	Blind with fixed length	Learned	CNN	Private	1019	Identification	maximum 6 months	Not specified	5 s
[20]	R-peak with fixed length	Learned	Neural network	Public	90	Verification	9 days for EDGIDDB	Not specified	Not specified
[4]	Blind with fixed length	Fiducial	Random forest	Public	1985	Verification	Not specified	1 m	3 s
[22]	R-peak with fixed length	Learned	RNN	Public	47–89	Identification	Not specified	18 heartbeats	18 heartbeats
[23]	Blind with fixed length	DWT	CNN	Public	18–47	Identification	Not specified	250 × 2-s segments	1 × 2-s segment
[18]	R-peak with fixed length	Learned	CNN	Public	52	Identification	Not specified	10 s	10 s