To test the robustness of each tool’s handwriting OCR, we anonymized a collection of samples from real medical office forms and sent the same ~1,000 values to each system. Real medical form data was chosen because it combines many challenging elements:
- Complex words
- Cramped space for entering values
- A high-stress, fast-paced environment when people are writing
In order to test Automation Hero’s new Context Aware OCR capabilities, the Hero Platform_ was also given a custom Knowledge Graph. This knowledge graph was created by the customer with more than 40,000 potential values for the form field we were testing.
Our patent-pending technology enables the AI to directly use this data when making a prediction, leveraging your team‘s stored knowledge and experience to automatically improve results.
If you want to know how our platform can improve your invoice processing experience get in touch to request a demo.
