Let’s say you’ve received a handwritten letter from a friend and want to edit it on your computer. You scan the letter and open your OCR programme. Here’s what happens:

1. First, the software analyses the layout of the letter. It recognises where the text and any drawings are located, remembers their position on the page and counts the paragraphs and special elements such as the date or signatures.
2. Then the tricky part begins. The software looks at the blocks of text and breaks them down into sentences. These sentences are then broken down into individual words and finally into letters.
3. The OCR programme has already learned patterns of letters and characters. Now it compares the scanned letters with these patterns. If the match is high, the algorithm says that this is most likely the correct letter. It is very accurate and can compare many patterns in a short time, so it can even distinguish between similar letters, such as a “c” and an “e”.
4. In this way, all letters and characters are gradually recognised. They are then put together to form words and returned to their original position in the letter. As soon as the software is finished, you can save the letter as a normal document and edit it as you wish. That’s it!