Improved Data Extraction Using Azure AI Vision
We enhanced the method of retrieving patient details from PDFs by transitioning from
AWS Textract to Azure AI Vision. Azure AI Vision’s ability to analyze images and read
text with prebuilt image tagging, text extraction with OCR, significantly improved the
accuracy and performance of the data extraction process.
Custom Model Training for Checkbox Recognition
A bespoke model was developed using Azure AI Vision to proficiently recognize ticked
and non-ticked checkboxes within images and scanned PDFs. This strategic approach
substantially elevated the accuracy in identifying relevant information, with the model
being trained to discern between marked and unmarked checkboxes.
Efficient Entity Identification through Key-Value Pairs
Employing Azure AI Vision, we systematically grouped detected checkboxes and
generated key-value pairs for efficient entity identification. This structured organization
of data simplified the interpretation and utilization of patient details, providing our
client with a more streamlined and effective output.
Tech Stack







