OCR Web

This is a web app which uses the Tesseract API for Optical Image Recognition. This app will be deployed heresoon.

The basic functionality of the application is demonstrated in this video

Usage Instructions

Prerequisites

• leptonica

• tesseract

• node

• npm

Instructions for contributing can be found -> CONTRIBUTING.md file.

Installing Dependencies

• Clone this repository .

• Run npm install in the root directory to install the dependencies.

• Change to public directory and build the Angular app.

• Run npm start from the root directory.

• The app is now available at http://localhost:3000

Tools and Frameworks used

Front end

• Bootstrap

• AngularJS

Server side

• Nodejs (with Express)

• Node Tesseract : Used as a wrapper for using the tesseract API's for the Node Platform

• Fred's ImageMagicks textcleaner Bash scripts : To enhancing the image and reducing noisefor better read.

• Tesseract OCR Engine

• Multer : For managing the uploaded image files (Node)

Limitations

• As of now, it is implemented to recognize only English characters.

• Though the Tesseract-OCR engine is powerful, there is a limitation to its performance.

Future additions

• Reduce noise in the uploaded image (Clean).
• Display the cleaned image in front-end.
• Use multipart formdata to upload files through Angular HTTP Request(and add other processing functionalities).
• Copy text to be implemented using ngClipboard angular directive.
• Retrieve text from multiple images in a single request.