Description
Welcome to Course “Intelligently Extract Text & Data from Document with OCR NER” !!!
In this course you will learn how to develop customized Named Entity Recognizer. The main idea of this course is to extract entities from the scanned documents like invoice, Business Card, Shipping Bill, Bill of Lading documents etc. However, for the sake of data privacy we restricted our views to Business Card. But you can use the framework explained to all kinds of financial documents. Below given is the curriculum we are following to develop the project.
To develop this project we will use two main technologies in data science are,
- Computer Vision
- Natural Language Processing
In Computer Vision module, we will scan the document, identify the location of text and finally extract text from the image. Then in Natural language processing, we will extract the entitles from the text and do necessary text cleaning and parse the entities form the text.
Python Libraries used in Computer Vision Module.
- OpenCV
- Numpy
- Pytesseract
Python Libraries used in Natural Language Processing
- Spacy
- Pandas
- Regular Expression
- String
As are combining two major technologies to develop the project, for the sake of easy to understand we divide the course into several stage of development.
Stage -1: We will setup the project by doing the necessary installations and requirements.
- Install Python
- Install Dependencies
Stage -2: We will do data preparation. That is we will extract text from images using Pytesseract and also do necessary cleaning.
- Gather Images
- Overview on Pytesseract
- Extract Text from all Image
- Clean and Prepare text
Stage -3: We will see how to label NER data using BIO tagging.
- Manually Labeling with BIO technique
- B – Beginning
- I – Inside
- O – Outside
Stage -4: We will further clean the text and preprocess the data for to train machine learning.
- Prepare Training Data for Spacy
- Convert data into spacy format
Stage -5: With the preprocess data we will train the Named Entity model.
- Configuring NER Model
- Train the model
Stage -6: We will predict the entitles using NER and model and create data pipeline for parsing text.
- Load Model
- Render and Serve with Displacy
- Draw Bounding Box on Image
- Parse Entitles from Text