traingry.py is the training code.
If you want the weight of the color image, choose net = DRFENet(channels=3,num_of_layers=17).
If you want the weight of the gray image, choose net = DRFENet(channels=1,num_of_layers=opt.num_of_layers).
Lines 224 and 226 of the code will automatically generate the .h5 file you need for your training
The gray dataset of train and val and test can be downloaded at the following link:
link：https://pan.baidu.com/s/1BzJKf236Fl1Sv7Q4_cyfyQ
Code：o83k
The RGB dataset of train and val and test can be downloaded at the following link:
link：https://pan.baidu.com/s/1eDq_lqksZI85__6jmZX6qA?pwd=3sib
Code：yo1b
test(gray).py is used to test the performance of the gray image model
test(RGB).py is used to test the performance of the RGB image model
Choose between them individually
net = DRFENet(channels=1)
net = DRFENet(channels=3)