Text Extraction From Various Vendor’s Receipts

Finance Platform
Text Abstration

Overview

Insurance firm wanted us to build an OCR platform which can extract information from any given receipts from any number of vendors.

Technology

  • Deep Learning
  • Neural Networks
  • Computer Vision
  • Image Processing

Language

  • Python

Key Technical Challenges:

  • Extracting information situated without any proper format from the images.
  • Mapping of parts synonyms with their original names in database without prior knowledge of it.

Business + Technical Points:

  • Effective extraction of words & numbers
  • Many images may have hand written words & numbers, they also must be extracted.
  • Effective mapping system for our needs.

Result:

  • OCR engine with much efficiency was built, integrating in their system.