Comparing pre-trained models for efficient leaf disease detection: a study on custom CNN

Journal of Electrical Systems and Information Technology

Table 10 Overview of the results

Model name	Accuracy	F1-Score	Macro-Avg	Precision	Recall	Trainable Parameter	Training time in seconds	Weighted average
Densenet201	0.997	0.997	0.997	0.997	0.997	18,824,010	3915.20	0.997
EfficientNetB3	0.998	0.998	0.998	0.998	0.998	11,185,721	6740.14	0.998
EfficientNetB4	0.999	0.999	0.999	0.999	0.999	18,142,569	9768.72	0.999
Inception ResnetV2	0.998	0.998	0.9988	0.998	0.998	54,738,922	5375.27	0.998
MobilenetV2	0.998	0.998	0.998	0.998	0.998	2,556,938	2265.5	0.998
ResNet152	0.999	0.998	0.999	0.998	0.998	58,906,250	4529.14	0.999
Resnet50	0.998	0.998	0.998	0.998	0.998	24,123,018	3358.5	0.998
Vgg16	0.981	0.997	0.981	0.999	0.997	14,850,634	4176.72	0.981
Xception	0.999	0.99	0.999	0.99	0.99	21,396,786	5683.83	0.999
LDDTA	0.979	0.975	0.98	0.976	0.975	184,890	1891.22	0.98