Credit Card Digits Recognition

Detects and recognizes credit card digits (name, number, valid date, type: visa/mastercard) using AI and CV.

Getting Started

To see how I collected the dataset and trained my model please read the README in 'dataset' directory

After you train your model save it as "models/darknet-yolov3_best.weights"

I coudln't attache the dataset because of private details of the cards holders

Prerequisites

python >= 3.6

python libraries:

• opencv-python (cv2)
• flask
• flask_restful
• jsonpickle
• numpy

Installing

Install libraries using pip3

pip3 install -r requirements.txt

Usage

1. Standalone application

python3 main.py <image_path> 

Option: image_path = examples/6.jpg or image_path = examples/8.jpg

Image should be credit card image with minimum background.

2. Rest API - running https server

python3 REST_api.py

Optional:

python3 REST_api.py --host <host_ip> --port <port_number>

Default argument:

--host = 0.0.0.0

--port = 5000

clinet url should be POST method and looks like this: http://<host_ip>:<port_number>/pic

Attach to the request credit card image that filmed from Simplify application.

Acknowledgment

Dataset:

• Friends, especially Guy Chriqui :)
• Imagaug

Training:

• AlexeyAB/Darknet
• Training YOLOv3

License

This project is licensed under the MIT License - see the LICENSE.md file for details

Authors

Shira Levy - Shira Star

"It is only with the heart that one can see rightly, what is essential is invisible to the eye" Le petit prince