An attempt to completely automate cheque verification and money transfers in banks to reduce human labour.
Using computer vision, siamese networks and optical character recognition(OCR) to extract and process data from bank cheques, thereby automating the process of cheque verification in Banks.
- HTML
- CSS
- VANILLA JAVASCRIPT
- NODEJS
- EXPRESS
- PYTHON
- TENSORFLOW
- KERAS
- TESSERACT
- SQLITE
- VISION API
Before we start the installation, first ensure that you have created a project on Google Cloud with Vision API enabled in it. To do so follow this link Quick Start Guide. Follow this guide till Install the client library section.
Ensure that your system have the following softwares installed:
You can download them by clicking on the link.
First download the Github Repository into your desired folder by using the following command.
git clone https://github.com/Blastoise/Cheque-Verification.git
Now install all of the required Python libraries required to run the code using requirements.txt given in cheque-verification folder.
To do so, execute the following command:
cd Cheque-Verification && cd cheque-verification && pip3 install -r requirements.txt
Now we need to install all the required packages mentioned in package.json under cheque-verification-server.
Execute the following command and all the required packages will be installed and a node_modules folder would be created in cheque-verification-server.
cd ../cheque-verification-server && npm install
You may need to give permission for running npm install
.
Now copy the mcr.trainedata from resources folder into tessdata folder which would located inside the Tesseract folder where you have installed Tesseract.
In Windows, in general it is located at:
C:\Users\USER\AppData\Local\Tesseract-OCR\tessdata\mcr.traineddata
In Linux and MacOs, in general it is located at:
/usr/share/tesseract-ocr/4.00/tessdata
To run the application, first go into cheque-verification-server folder and execute the index.js file using NodeJS.
To do so, execute the following (Assuming we are already in cheque-verification-server folder):
node index.js
This will start the server at PORT 8000.
Note: Keep all the cheques inside cheque_images under cheque-verification folder.
-
For downloading the Signature Data on which we trained, you can follow the link below: https://cedar.buffalo.edu/NIJ/data/signatures.rar
-
Github Repo for mcr.tessdata: https://github.com/BigPino67/Tesseract-MICR-OCR