Table 1 Comparative performance of PLSR, CNN, and DR-CNN models on Quality indicators.

From: Quality prediction of air-cured cigar tobacco leaf using region-based neural networks combined with visible and near-infrared hyperspectral imaging

Model	Quality Indicator	Source Dataset	RMSEC	RMSEV	RMSEP	R2 (Test)
PLSR	Moisture	NC + IR	2.48	4.19	4.53	0.74
PLSR	Moisture	NC + RL + IR	3.09	3.51	3.85	0.76
PLSR	Moisture	ALL	3.82	3.20	3.40	0.78
CNN	Moisture	NC + IR	2.34	3.91	4.20	0.75
CNN	Moisture	NC + RL + IR	3.04	3.42	3.60	0.80
CNN	Moisture	ALL	3.72	3.11	3.30	0.79
DR-CNN	Moisture	NC + IR	2.08	3.83	3.96	0.77
DR-CNN	Moisture	NC + RL + IR	2.77	3.20	3.42	0.81
DR-CNN	Moisture	ALL	3.53	2.91	3.11	0.81
PLSR	Chlorophyll	RL + BL	0.98	1.00	0.89	0.51
PLSR	Chlorophyll	RL + BL + UV	0.51	1.12	0.91	0.54
PLSR	Chlorophyll	ALL	0.51	0.90	1.38	0.55
CNN	Chlorophyll	RL + BL	0.66	0.62	0.79	0.61
CNN	Chlorophyll	RL + BL + UV	0.59	0.69	1.04	0.69
CNN	Chlorophyll	ALL	0.53	0.73	0.97	0.63
DR-CNN	Chlorophyll	RL + BL	0.50	0.79	0.83	0.73
DR-CNN	Chlorophyll	RL + BL + UV	0.44	0.64	0.82	0.67
DR-CNN	Chlorophyll	ALL	0.69	0.50	0.88	0.76
PLSR	Nitrogen	VL + NC	0.15	0.20	0.22	0.78
PLSR	Nitrogen	VL + GL + NC	0.18	0.16	0.18	0.81
PLSR	Nitrogen	ALL	0.20	0.16	0.16	0.85
CNN	Nitrogen	VL + NC	0.13	0.22	0.25	0.77
CNN	Nitrogen	VL + GL + NC	0.15	0.18	0.20	0.79
CNN	Nitrogen	ALL	0.18	0.15	0.18	0.83
DR-CNN	Nitrogen	VL + NC	0.11	0.18	0.20	0.81
DR-CNN	Nitrogen	VL + GL + NC	0.13	0.15	0.17	0.84
DR-CNN	Nitrogen	ALL	0.16	0.13	0.15	0.86
PLSR	Sugar	VL + NC	0.15	0.16	0.17	0.71
PLSR	Sugar	VL + BL + NC	0.14	0.15	0.16	0.72
PLSR	Sugar	ALL	0.13	0.14	0.15	0.73
CNN	Sugar	VL + NC	0.14	0.15	0.16	0.73
CNN	Sugar	VL + BL + NC	0.13	0.14	0.15	0.74
CNN	Sugar	ALL	0.12	0.13	0.14	0.75
DR-CNN	Sugar	VL + NC	0.14	0.15	0.16	0.73
DR-CNN	Sugar	VL + BL + NC	0.13	0.14	0.15	0.74
DR-CNN	Sugar	ALL	0.12	0.13	0.14	0.76

Back to article page

Table 1 Comparative performance of PLSR, CNN, and DR-CNN models on Quality indicators.

Search

Quick links