Optical Character Recognition (OCR) overview and examples
2011/04/12 Leave a comment
Today we are reviewing the Optical Character Recognition technology, which our software and application use.
First of all let’s start with the definition:
[OCR]“is the mechanical or electronic translation of scanned images of handwritten, typewritten or printed text into machine-encoded text.” from wikipedia
The nature of this technology is to simplify the entry of document into a bigger process. The goal is to have a digital version of a paper document and avoid the manual entry of characters.
Nowadays, main corporations already integrate this technology into the company’s global process, from billing to document archive.
(1) Scan or picture capture of the document: a scanner or a camera capture the paper document.
(2) The OCR engine extract the text from the image
(3) The text recognized is send to be processed
Optical Character Recognition is not perfect, as their are some factors that change the recognition quality:
- Resolution: High-resolution files work best as more details helps the recognition
- Orientation: Most engine recognizes only documents with horizontal left-to-right text.
- Languages, fonts: common fonts works better. Handwriting fonts tends to be not well recognized. The language defines also the set of character that can be recognize, some characters being specific to a language.
- Image quality: Sharp images with good lighting and clear contrasts will work best. Blurry image will reduce the quality of the recognition.
The OCR technology compare the patterns recognized on the picture to the pattern of fonts. So as if more and more fonts are well recognized this days, their are still some fonts or text orientation that are most likely to be not recognized.
The current average character recognition performance is around 95%.
3. Common applications
Apart from corporate usages, like digitalizing document for archive, OCR is also use in public services.
- ID card and passport scanner
- Car plate scanning for parkings
- Document reader for disabled people
4. Mobile applications
Smart phones being now in everybody’s pocket, OCR simplify the entry of text. You can now enter a contact info directly into your mobile phone simply by taking a picture of the contact’s business card. You can translate menu or signs when you’re traveling. You can scan a document and integrate it to your company data process while on the move without a computer.
example of application using OCR technology: Scancard-Business Card Reader for iPhone