DMS Tips

OCR - how companies benefit from optical character recognition

OCR is a technology for optical character recognition of digital images. When physical mail is digitised, for example, the resulting files can be fully searched thanks to full-text search. We show how OCR works and how this technology can support you.

‍

What does OCR mean?

What is OCR? The abbreviation OCR stands for optical character recognition and refers to a technology for text recognition in digital images.

So what do things like invoices have to do with it? An invoice isn't an image, it's a text document. True – and yet, not quite. In the Caya OCR process, we first scan documents as image files. Printed documents or handwritten text are read and, as individual characters are recognised, converted into a digital format. The original document can be text (like an invoice) or an image containing text (e.g. in advertising graphics, flyers, or brochures). The source document is first captured as a digital image and then read during the OCR process. The OCR process for document digitisation and data capture is also used for passports, bank statements, and business card details, among other things.

Optical character recognition and OCR software are also fields of research within artificial intelligence (AI), feature matching, and pattern recognition. When capturing data, such as from paper letters, OCR software analyses the document structure. It divides the document into various elements such as sender, subject line, and body text. This involves global structure recognition as part of a layout analysis. This can distinguish text blocks from graphic elements, allowing it to recognise both line structures and individual characters. The program stores the location of each piece of content.

During text conversion, individual extracted lines of text are broken down into words and then into individual letters. An image created by a scan consists of a collection of pixels, as does every letter. For optical character recognition, these pixels are compared with a series of patterns by algorithms (pattern matching). This allows individual letters to be defined across a wide variety of typefaces. A text layer containing computer-interpretable characters (ASCII characters) is then placed over the scanned image file.
‍

OCR scan – accurate reading of text information from images

The typography/font of letters, as well as the handwriting in written letters, varies greatly depending on the sender. When performing OCR on an image file, the OCR software must reliably read the available information correctly. To this end, an additional system for context recognition, known as "ICR" (intelligent context recognition), is used. ICR assists by allowing incorrectly recognised characters to be corrected based on context. What does this mean? For example, reading the number "8" instead of the capital letter "B" would be a reading error. Without ICR context recognition, "Bus" could easily become "8us". ICR ensures the appropriate correction here. Alternatively, it also ensures that alphanumeric terms like "8ter" are not converted due to context.

It is also important to note that error correction improves with an increasing number of digital text conversions with every scan. This is where machine learning comes in. This happens because OCR technology remembers where standardised and recurring content (e.g. a letterhead or an invoice field) is located. As a result, text conversion is quickly optimised with every new scan. Another advantage of documents digitised by OCR software is that a newly created PDF file can subsequently be searched using full-text search. This is also the case in the Caya Document Cockpit. In the document centre, you can search documents for specific terms or passages of text.

‍

Scan as OCR – no more manual typing

Digitising documents via OCR scan significantly increases business productivity. Automated data processing makes your workflow much easier compared to manually entering the information contained in documents. Documents are often searchable for all terms they contain. When linked with smart accounting software, for example, manual typing into other systems is often no longer necessary. With the appropriate integration, data transfer to other software happens automatically.

In addition, the use of an OCR solution also increases internal company security standards. Costly and insecure local data storage becomes unnecessary when you choose a GoBD-certified provider like Caya as your document centre.

‍

Automatically categorised document filing thanks to OCR

Important information such as the sender or subject is automatically captured during the OCR process. This also applies to invoice processing, for data such as the invoice amount, payee, IBAN, BIC, or payment reference.

Caya automatically classifies incoming documents. This makes it very easy to filter for invoices, for example. An incoming invoice is also recognised as such (invoice recognition) and categorised accordingly. Documents can also be automatically forwarded to the relevant department. Scanning and archiving are made much easier during invoice processing by an automatically categorised scan. On the way to a paperless office, the tedious manual filing of (digital) documents thus becomes a thing of the past.

‍In general, OCR offers the following key benefits for the paperless office:

  • Time- and location-independent document access
  • Easier access to information for others
  • Scan as PDF fully searchable thanks to full-text search
  • Enhanced editing functions (including copy & paste option)
  • Automatic mail categorisation (tagging)
  • Digital, audit-proof document filing in the Caya Document Cockpit
  • Possibilities to connect automated processes to digitised documents

‍

OCR as the foundation for automated workflows

Physical mail is usually the first step in document management process chains. Caya digitises your incoming post. To do this, we work with Deutsche Post and PIN AG to redirect your mail to one of our scanning centres. We then digitise your post using a highly automated process. Digitising incoming mail is the essential first step in integrating paper documents into digital workflows and accelerating their processing.

Invoiced data read from companies can be automatically transferred to the relevant accounting tools via integrations. Invoices are scanned and automatically inserted into accounting tool interfaces via APIs. Administrative tasks such as manual data entry are eliminated. When linked to the appropriate billing software, the data read is automatically entered into the programmes. As a rule, all that remains is to check and confirm. Responsible staff only receive all the data for a final decision during the monitoring process (decision-ready). This reduces the burden, allowing you to focus on value-adding activities. Furthermore, invoices can be paid online via your Caya account with just a few clicks. Payment is made as a SEPA transfer directly from your account.

Paper plane

Would you like to find out more?