OCR Software
SUPPORT AND MAINTENANCE
Lancer:
<= 2 resources OR <=3 weeks effort OR <=$3000 USD
Growth:2 - 4 resources AND upto 4 months OR
upto $30000 USD
Enterprise:Everything else
(Development)
Project Description
A Switzerland-based digital solutions provider aimed to automate PDF-based data extraction for faster inventory mapping. With a growing need for structured article pricing data, the client sought a robust backend module to streamline extraction, classification, and integration. The solution needed to support real-time processing through a web-based interface for GraphQL-connected systems.
Scope Of Work
The client sought a solution with Oodles for a system to extract article numbers and prices from PDF files and fetch related results via GraphQL API. The goal was to replace manual data collection with a scalable module. Key areas of work included PDF parsing, OCR integration, NER model execution, GraphQL connectivity, and structured export generation.
Our Solution
To address the client’s need for automation, a Django-based REST API system was built to handle PDF-to-text conversion, entity recognition, and result generation. Here's how it worked:
- Image Conversion Module: PDF files were converted into JPGs using pdf2img for improved OCR accuracy.
- Text Extraction: Google Tesseract OCR processed the images to extract raw textual content, saved in intermediate .txt files.
- Entity Recognition: A spaCy-based Named Entity Recognition (NER) model was trained to extract specific fields from the text.
- Data Mapping & Export: Extracted data was sent through a GraphQL API to fetch the updated article details.
Share Your Requirements