Table 1 Performance analysis of the current literature in comparison to the tenfold average accuracy of the examined SVM-based and ANN models. Bold font indicates the best performance.

	State-of-the-art	Method	Input features	ACC (%)
Literature	Kather et al.¹⁰	SVM	GLCM, LBP, Gabor	87.4
	Cascianelli et al.¹¹	NNC/VGG-based “off-the-shelf”	LBP, Texture Spectrum, PCA / Deep Features	79.60/84.00
	Sarkar et al.¹²	SDL	Local Gabor filtering	73.7

	WPT family	Levels	# Features	ACC (avg ± std%)
Baseline SVM	rbio	1	155	84.04 ± 1.98
	bior	2	155	83.22 ± 2.34
	coif	2	140	83.21 ± 1.24
	sym	1	135	83.04 ± 1.63
	rbio	2	532	82.76 ± 1.67
	bior	2	532	82.18 ± 1.35
	coif	2	532	82.16 ± 1.57
	db	2	532	82.10 ± 1.68

Quick links

Search