Hybrid Techniques for Diagnosing Endoscopy Images for Early Detection of Gastrointestinal Disease Based on Fusion Features

Table 1

Adjust hyperparameters of VGG16 and DenseNet121 models.

Options	VGG16	DenseNet
Training options	Adam	Adam
Mini batch size	55	30
Max epochs	35	22
Initial learn rate	0.0001	0.0003
Validation frequency	50	50
Execution environment	4 GB GPU NVIDIA	4 GB GPU NVIDIA