Table 1.

Data distributions for the datasets used in our comparative experiments.

Name	Class	Diseases	Training set size	Validation set size	Testing set size
Wheat	0	health	528	65	59
	1	rust	673	83	77
	2	mildew	282	34	32
	3	smut	674	83	75
	4	root rot	381	46	41
	5	scab	391	48	45
	6	leaf spot	378	47	45
Coffee	0	healthy red	353	43	39
	1	spider mite	136	16	15
	2	rust	324	39	35
Rice	0	healthy	407	50	44
	1	unhealthy	413	50	43