Can you please explain working process of it. Coz I did not get some things.

I didn't understand what is the model trying to achieve. Which class of signatures as mentioned in the "ReadMe" section.

Could you please elaborate.

The training images and the testing images are not visible, and the CSV output file is not clear.

if the scores in the output file are the probability that a specific signature matches a specific character, then the accuracy will be very very low!

Hi,

Please help me to dump the train result in pickle file for code re-usability.
Thanks in advance.

Thanks,
Bharath J