FLask app for Named-Entity-recognition

About

Simple app for extracting entities like (Name , Designation , PAN , UAN no. , Salary , PF , Date of joining etc) from payslips.

How it Looks

Prediction page

Steps for building app

Step 1. Collected payslips data for model training.

Step 2. Used Pytesseract-OCR for extracting text from payslips followed by text pre-processing.

Step 3. Annotated data with online NER annotator tool.

Step 4. Trained a Custom NER model using spacy.

Step 5. Built a web app using Flask package which returns extracted entities from uploaded document.

How To Use

Run it locally

Step 1. Clone the repository

Step 2. Pip Install all dependencies in requirements.txt in your conda environment (You may run pip install -r requirements.txt in your shell )

Step 3. Run python Flask_app.py in cmd prompt