Table 1.

Summary of all the identified publications on ECG quality assessment techniques. Note that FB refers to feature-based and NFB refers to non-feature-based. N/R refers to not received, meaning that the value was not mentioned in the article. The values have been averaged for studies that reported multiple results. * Averaged value.

Name First Author and Year	Type	Data	Leads	Cross Validation	Sensitivity (%)	Specificity (%)	AUC (%)	Accuracy (%)
Fu et al. (2021) [5]	FB	Private dataset	1	N/R	97.41	86.50	N/R	95.63
Huerta et al. (2021) [6]	NFB	CinC17 [7]	1	N/R	90.00	91.00	N/R	90.00
Liu et al. (2021) [8]	NFB	CinC11 [4]	12	N/R	97.67	84.73	N/R	93.09
Xu et al. (2021) [9]	FB	Private dataset	1	N/R	N/R	N/R	N/R	95.23 *
Huerta et al. (2020) [10]	NFB	CinC17 [7]	1	N/R	90.37 *	93.00 *	N/R	87.77 *
Hermawan et al. (2019) [11]	FB	CinC11 [4]	1 + 5	N/R	92.00 *	67.00 *	N/R	85.75
Moeyersons et al. (2019) [12]	FB	CinC17 [7], NSTDB [13]	1	3-fold	98.37	96.43	99.67 *	N/R
Ansari et al. (2018) [14]	NFB	MITDB [15], NSTDB [13]	2	N/R	N/R	N/R	96.41	N/R
Liu et al. (2018) [16]	FB	CinC11 [4], CinC14 [17], TELE ECG [18], Private dataset	12	10-fold	89.08 *	88.15 *	N/R	97.15
Yaghmaie et al. (2018) [19]	FB	CinC17 [7], NSTDB [13], MITDB [15]	1	5-fold	96.20	97.70	93.18	96.90
Zhang et al. (2018) [20]	NFB	CinC11 [4], CCDD [21]	12	N/R	N/R	N/R	N/R	96.45
Athif et al. (2017) [22]	FB	CinC11 [4]	12	5-fold	91.20	91.60	N/R	91.10
Satija et al. (2017) [23]	FB	CinC11 [4], MITDB [15]	N/R	N/R	N/R	96.71	N/R	N/R
Taji et al. (2017) [24]	NFB + FB	MITDB [15], NSTDB [13]	1	4-fold	98.20	98.30	N/R	97.20
Xia et al. (2017) [25]	FB	CinC11 [4]	12	N/R	N/R	N/R	N/R	91.60
Orphanidou et al. (2016) [26]	FB	Private dataset	1	N/R	92.50 *	98.00 *	N/R	95.00 *
Morgado et al. (2015) [27]	FB	CinC11 [4]	12	10-fold	86.08 *	89.05 *	91.25 *	92.25
Li et al. (2014) [28]	FB	CinC11 [4], NSTDB [13], MITDB [15]	12	N/R	N/R	N/R	N/R	96.51 *
Clifford et al. (2012) [29]	NFB + FB	CinC11 [4], NSTDB [13]	1 + 12	N/R	89.00	99.00	N/R	97.00