In your paper, there are 20 examples of data enhancements. In the process of training, change 20 enhancements for each image?
196 samples->license-plate-detection.py
196 samples-> 196*20(augment sample)->license-plate-detection.py
Do you choose process 1 or 2
I choose process 1 to use 105 samples in cars ,but His effect is very poor
You use 200 samples to train LP detector. is it too few?
augment sample