Expense Tracker

Team Name: GroupTwo
Team: Shubham Bisht
Shadrak Gurupnor
Mukund Choudhary

Problem Statement:

Key Information Extraction from Scanned Receipts: The aim of this task is to extract texts of a number of key fields from given receipts, and save the texts for each receipt image in a json file.

    Step 1: Extracting all the text from given Invoice Image. (Text Region Detection, OCR)
    Step 2: Recognising Key information from the text like Store Name, Address, Total Amount etc using Bi-directional LSTM based approach

Usage Guide

This repository contains three directories. Inside each folder there are documentations of the method we adopted and guide of usage.

• Folder 1 - Bill Label Tool: A GUI tool for efficiently labelling various entities in huge dataset of invoice images.
• Folder 2 - ML: Bi-LSTM Approach for Character-wise classification of invoice text into key entities.
• Folder 3 - Web UI: A user interface for uploading Invoices and receiving results in a HTML form

• Folder 4 - Android UI: Android interface for uploading Invoices, receiving results and to watch previous Bill uploads. Find the Android Repo here

Approach

• Task 1 - Scanned Receipt OCR: Tesseract OCR
• Task 2 - Key Information Extraction: Character-wise classification with Bi-LSTM

For the information extraction task, each image in the dataset is annotated with a text file with format shown below:

{
  "company": "STARBUCKS STORE #10208",
  "address": "11302 EUCLID AVENUE, CLEVELAND, OH (216) 229-0749",
  "date": "14/03/2015",
  "invoice id":"5628391",
  "total": "4.95",
  "items": "WhiteMochaV"
}

Results