OCR

OCR software comes in many different types, which vary in price range based on their features, speed and accuracy. For example, you can get a freeware like SimpleOCR that will serve you in no time, but it will only be able to convert BMP, JPG and TIF images of English or French text into plain text documents of TXT or DOC format, one page at a time.

On the other hand, you can invest a few hundred dollars in a Batch OCR or even Server OCR software that can view certain folders for incoming documents in different image formats and languages and then automatically make exact copies of all the pages in them in a format of your choice.

You can also find Desktop OCR software, which will bridge the price gap and include many of the features of the Corporate editions, but still requires some user input during conversion. 

Accuracy of OCR Software

Although some OCR engines are better than others, no software can guarantee 100% accuracy. This is because other factors come into play, including scan quality. Recognition software will not be able to do its job if the scanner does not properly digitize the page.

It is recommended to scan at a resolution of 300 dpi for best results. Black and white (Bitonal) is preferred over greyscale or color mode, and although most modern scanners are configured fairly well out of the box, you can adjust your brightness and contrast settings for your specific documents.

OCR software is also limited in what it can recognize. Most OCR software is only designed to recognize machine-printed text, as opposed to handwriting. While there is ICR software that can recognize handwritten information, these are usually enterprise-level solutions for forms processing, rather than full page recognition.

Similarly, most OCR software are only able to convert traditional machine fonts, not italic scripts or calligraphy. There are many fonts available and OCR engines rely on commonly used, separated font shapes to recognize the text, so fonts that are unusual or blend together will not be recognized.

Differences in OCR Software

The main features that distinguish OCR software are:

• Character recognition accuracy
• Correction accuracy page layout
• Support for languages
• User interface design
• Output file formats (Word, Excel, PDF, eBook, etc.)
• OCR speed and support for multi-core CPUs
• Batch processing modes
• Advanced PDF encoding or compression
• Special features for niche projects

Because of the infinite combinations of document types, OCR engines, project requirements and special features, one engine may perform better with your particular documents than with another.

Despite the benefits that e-invoicing has, its use is still anything but commonplace. Indeed, e-invoicing is not only an internal matter, but depends on the cooperation of suppliers. Moreover, there is no single e-invoice standard, but many different XML-based standards, such as HR-XML, SETU, various UBL standards, Finvoice. It is therefore a matter of converting incoming invoices from suppliers to the desired standard.

In this, there are several solutions that vary in degree of relief. Incoming paper invoices, for example, can be scanned yourself and then translated online into a validated data file (XML). You can do that translation yourself, validating the scanned document (accounts payable data) and encoding it (general ledger accounts), then sending the validated data file into a procuration workflow for approval by budget holders.