hand-writting-detection

Sherwin is insane...

Description

Looks at two different signature samples, and tells whether or not the handwriting is a forgery or not

Usage

Can fill the training-data folder with whatever letter images as long as the names are in the correct format, either:

1. [letter]xxx.filetype where [letter] represents which letter it's supposed to be, xxx are 3 distinct digits, and the filetype is either .png or .jpg
2. [letter] (X).filetype where [letter] represents the letter it's supposed to be, X is any number, and filetype is either .png or .jpg

Run setup.py to train the model based on the given training data. It will output a graph for accuracy rate

Run main.py for the machine to use the trained model to identify images of letters.

Dependancies

• tensorflow: machine learning library
• keras: deep learning, nerual network APIs
• sci-kit image: image processing (used to resize image)
• numpy: array processing (images are stored in an array)
• matplotlib: plotting graphs
• opencv: image processing

Have fun!