Credit Card Number Reader

This project was a part of my online course "Computer Vision Opencv4 with python and deep learning".

As most Credits cards use the fonts Ferrington 7B and OCR - B fonts, we made use of those credit cards with above fonts for testing and training.

The steps used in arriving the solution is given below

Enter the 2 font sized images commonly used
Convert into Grayscale
Create test and train data directories
Create different versions of the initially entered images (different versions for both the fonts)
Preprocessing the data (Introducing noise, rescaling, upscaling/downscaling
Now transfer these images into the directories created
Used LeNet CNN for training the numbers
Now test it on the test dataset that we created
Now we test this on the actual credit cards
We made use of the Canny edge detector in order to find the edges of the card
Then we made us of 4 point transform to align the card to extract the card numbers
Now once this is done, we use findContours function to isolate the digits from the card
Then we pass this onto a classifier.

Programming language: Python with OpenCV