Handwriting Recognition and OCR

Automation of business processes sometimes requires extracting contents from scanned documents — images and pdfs. These documents may contain both machine generated texts and handwritten contents. Optical Character Recognition (OCR) is used to extract contents from scanned documents, which are typically converted into images when scanned.

Most OCR technologies available in the market can read machine-generated texts at some level of accuracy but they fail to extract handwritten contents. Our deep learning and computer vision-based trained models can read handwritings in addition to reading typed texts at a high level of accuracy.

We also provide an intuitive UI-based system to train your own models based on your unique handwriting sets — especially older documents.

In addition, you can train your own models to recognize non-English languages, such as Arabic, Chinese, Hindi, and more.

How It Works

  • Signup for a Accure Momentum account and request the API key
  • Make a call to the Restful API by sending a scanned document along with the API key.
  • The API returns a response in JSON format that contains text with line number and the bounding box references.

If you are interested in trying the handwriting recognition API, signup to get the API key and documentation.