Procedures:

#1. Pre-handle the images using opencv #2. Conv and Pool using vgg16 #3. Detect edges of characters #4. identify characters #5. train #6. test