A Crop Leaf Disease Image Recognition Method Based on Bilinear Residual Networks

Table 9

Model computational complexity performance on different crop datasets.

Dataset	Training time (seconds/per epoch)		Recognition time (milliseconds/per image)		Model size (MB)
	ResNet-18	DIR-BiRN	ResNet-18	DIR-BiRN	ResNet-18	DIR-BiRN
Apple	4.0889	4.5587	0.1899	0.2045	12.23	11.18
Corn	4.9937	5.0733	0.1870	0.2091	12.23	11.18
Grape	5.3242	5.5607	0.1844	0.2065	12.23	11.18
Potato	2.8287	2.8495	0.1844	0.2052	11.96	11.18
Tomato	24.0697	24.2114	0.2015	0.2173	13.80	11.19
Combined	44.3843	44.6054	0.2012	0.2183	17.73	11.19