A Hemolysis Image Detection Method Based on GAN-CNN-ELM

Table 3

Model comparison.

Model	Features (number of input features)	Total feature extraction time (unit: s)	Duration of training (unit: s)	Total time (unit: s)	Accuracy (%)
GAN-CNN	—	—	—	5153.56	97.78
GAN-ELM	15360	1250.51	3.96	1254.47	53.78
GAN-ELM-L1	15360	1250.51	3.11	1253.62	59.33
GAN-SVM	15360	1250.51	14.83	1265.34	54.61
GAN-CNN-SVM	15360	1250.51	13.47	1263.98	98.61
CNN-ELM	15360	1085.28	3.41	1088.69	97.97
GAN-CNN-ELM	15360	1250.51	3.44	1253.95	98.77