Table 7. The accuracies performance of BHCNet-6 for the multi-classification.

References	Methods	40×	100×	200×	400×
Chan et al. (2016) [38]	SVM	55.6	-	-	-
Bardou et al. (2018) [4]	CNN	86.34	84.00	79.93	79.74
	CNN + Augmented	83.79	84.48	80.83	81.03
	SVM	82.89	80.94	79.44	77.94
	Ensemble CNN model	88.23	84.64	83.31	83.98
	BoW/DSIFT	66.72	69.06	62.42	52.75
	BoW/SURF	41.80	38.56	49.75	38.67
	LLC/DSIFT	60.58	57.44	70.00	46.96
	LLC/SURF	80.37	63.84	74.54	54.70
Present Work	BHCNet-6 + ERF	94.43 ± 0.28	94.45 ± 0.15	92.27 ± 0.08	91.15 ± 0.43