open close
calling ICO-LUX +49 3641 92 90 988 writing an e-mail info@ico-lux.com

ICO.Match - high-calibre document classification

To keep up with the times, it is important not only to automate document processing, but to take it to a new level of contextualisation. On the one hand, our ICO.Match technology is at the forefront of our intelligent document forensics ICO.Fraud. On the other hand, it is also used by many customers as a stand-alone solution for extracting information from documents (IDP) due to its outstanding text recognition and matching rates - successfully, for example, in the major area of energy certificate capture in accordance with the latest EU directives. Another major advantage of ICO.Match is that the document verification software ICO.Fraud can be seamlessly integrated in the event of fraud involving forged documents. But what exactly can ICO.Match do and what steps are behind it?

Benefits of ICO.Match

Structured data instead of data clutter

Optimised readout results make subsequent corrections obsolete. Structured data increases the data maturity of utilising companies.

Comprehensible information extraction

Unlike conventional OCR software, ICO.Match uses the layout to recognise content. This gives characters more context and turns them into relevant information and structured data.

Automatic document classification

ICO.Match accurately identifies and categorises submitted documents and classifies them into relevant, customisable document classes.

Faster, fairer decisions

ICO.Match accelerates the workflow and is ‘turnkey’ without the need to train algorithms. Processes become more efficient.

Data protection compliance

ICO-LUX stores neither documents to be extracted nor extracted information. Data processing takes place in certified data centres in Germany.

Integration and expansion

ICO.Match can be seamlessly integrated into existing system landscapes and, with ICO.Fraud, can be expanded to include high-calibre fraud prevention with minimal effort.

Countless documents of dubious image quality 

Anyone who has ever applied for a loan from a bank will know that pages and pages of documents are required for submission. Even insurance payouts do not work without the corresponding invoices. The situation is similar when importing goods from abroad, concluding rental agreements, buying property or applying for social benefits such as housing benefit. Paper documents are often still in circulation at these times, which are then photographed by non-professionals with smartphones, and not all documents that are then presented contain the required information. Above all, however, many of the documents that form the basis for important decisions are of poor quality and therefore elude automation efforts. ICO.Match image recognition and document processing helps to read documents in record time. On the one hand, this improves the customer experience due to faster feedback, and on the other hand, process costs can be saved as documents no longer have to be processed manually.

Smart vision technology as the basis for analyses

As a multi-award-winning technology, ICO.Match goes far beyond what was previously possible in the field of OCR - Optical Character Recognition. It can cope with poor quality documents, delivers fast results so that integration into automated application processes is possible and it converts content directly into structured data. This means that it enables in-depth analyses and automation processes, and these many steps in document processing can be automated: Specifically, skewed perspectives are corrected, any image noise is removed and the layout is reproduced in the form of tables and font positions. Everything recognisable is assigned to patterns and so the document can be contextualised in its document class with all its specifics. 
 

Requirements for the use of ICO.Match as on-premise software or SaaS

  • The documents to be analysed should be available in digitised form (PDF, scan or smartphone photo, at least 150 dpi)
  • Documents must contain a high proportion of printed text
  • For on-premise use, it is recommended to process at least 10,000 documents per day or a total document volume of one million documents

Why ICO.Match is redefining document processing

It doesn't take magic to turn a bad photo of a crumpled bank transfer confirmation or a poor scan of a coffee-stained copy of a rental agreement into a machine-readable document. But it does require high-performance document processing. With its technology, ICO.Match puts various providers of IDP software in the shade. This is because Optical Character Recognition all too often adds up recognised characters in long strings. AI-based IDP solutions initially require immense amounts of data for training and cannot be easily customised for special cases. Position-based data extraction only works for highly standardised documents and requires manual training.

ICO.Match, however, succeeds in reconstructing the meaning of the entire document in many small steps: document class, layout, text information, numbers - all of this flows into the algorithms for information extraction and enables extraction rates that outperform providers of IDP or input management solutions. ICO.Match takes the approach of capturing as much data as possible from the documents - and not just extracting individual pieces of information that are essential for certain business transactions. This enables companies to build the perfect database for increasing their own corporate data maturity and connect ICO.Fraud to optimise their fraud prevention.

Any questions? 
Let's talk.

Jan Franke

+49 3641 92 90 988 jf@ico-lux.com