How It Works

Handwriting-recognition tech

The ability for a piece of software to recognise handwriting is a natural extension of an older concept: optical character recognition (OCR). This was conceived around a century ago, when a machine was developed to aid telegraphy that read characters and translated them into code.

Conceptually, modern handwriting recognition (HWR) performs the same conversion, but with an additional process. The printed or handwritten document is first scanned, or written onto a touchscreen mobile or tablet device. The HWR app then separates each character and – using a pre-programmed bank of algorithms – matches it to what it thinks is the most likely letter on a database. Modern HWR software uses context to help decide one letter from the next and some programs can even ‘learn’ from reading a user’s writing over time, increasing efficiency.

Finally, the software creates a digital output, which can be read by any device and then replicated as editable text.

  • Peter Hackett

    Thanks for this post. I have a question for this excelent theme. An Enterprice need to do massive process of tickets data. This tickets have two class of data to need being to process it in a short form for cature the data, for example: Name: , Location: , Destiny: write in print data and the important handwritten response
    for being recognized. My question is if exist a machine tecnology for process this tickets, to take the two class of data and storage in a database to process there? Thaks for your responses, web links examples or tech distributors.