Client Story:

AI-based Oncology Scribing

The Client

The client, an oncology-focussed technology provider, implemented OCR-based image-to-digital scribing, with great success.

Key Challenges

System needed to:

  • Interface with legacy file delivery mechanisms such as Fax;
  • Communicate with file servers containing hand-written documents;
  • Scribe the scanned documents and translate the text to digital formats;
  • Create regression models for trained and untrained data

Our Solution

  • OCR technology for the reading interfaces;
  • Apache NiFi and Apache Kafka for the data ingestion;
  • NLP-NLTK for the scribing element.
  • The team set small achievable goals and hit MVP increments.
  • The team delivered MVPs in a manner which were production-usable.

Key Results Achieved

  • System now scribes hundreds of thousands of hand-written images and PDFs.
  • Complex business intelligence, data mapping and matching algorithms in place.
  • Regression models (trained and untrained) continue to improve the efficacy of the system.
  • NLP-NLTK effectively uncovered the hand-written notes for analytics.