For over 90% automatic OCR validation
Error-tolerant validation of OCR-captured forms using reference databases. Accurate even with poor read results. Includes time-saving manual follow-up support by offering intelligent search suggestions in real time.
- Supplements existing OCR for automated validation rates of over 90%
- Significantly reduces manual post-processing time & therefore process costs
- Simultaneously increases usability & manual post-processing quality
- All thanks to extremely fast & accurate matching, validation & searching
- Highly efficient even with high volumes & huge identity databases
- Intelligent, error-tolerant, language-independent & fully cross-field
- No premature use of restrictive keys or rules, although these are also available
- Can be connected to almost any OCR system in under a day thanks to powerful APIs
- Easy to connect to common ECM systems such as Top Image Systems, Readsoft & Captiva
Intelligent validation of OCR read results
exorbyte OCR Validate specializes in validating forms captured via OCR, e.g. invoices and applications, against a reference database such as customer lists and mail data. This allows you to get much more from your OCR. You can also achieve automation rates of over 90%, and reduce the time and cost of manual post-processing thanks to instant, intelligent suggestions.
This is all made possible by the underlying identity resolution engine, which factors in all possible similarities between characters intelligently and in real time, instead of merely relying on predefined rules that are based on statistical probability errors.
The core technology with FlexForm functionality also allows fields containing contextual errors to be validly handled. For example, MatchMaker ‘understands’ when a street has been entered in a name field.
OCR Validate takes into account the particular demands of read results produced by optical character recognition: these may include reject characters or transpose characters such as “U” and “V”. The hit rates for database matching are still high and solid when read results or queries are extremely corrupted. Even badly read and unrecognized characters, misspellings and incomplete data will be detected. Thanks to OCR Validate, a read result such as
“Schierner Antur – Salzbuche Slr. |5 – 60828 Frankturt”
will still be validly corrected and reliably identify
“Artur Schiemer – Sulzbacher Str. 15 – 60326 Frankfurt”
Terms with related content and the context (e.g. file number or customer ID) are also taken into account. Only this unique, complete matching method and the highest matching quality allows even OCR read results with enormous errors to be reliably corrected and validated.
Made for Big Players
New office in Karlsruhe
New office in Karlsruhe exorbyte is now represented with a new office in Karlsruhe. T…
Gartner positions exorbyte in the Magic Qua
Gartner positions exorbyte in the Magic Quadrant for Enterprise Search Search and mat…
Next Level Information Management with exor
To further optimize and automate the information management the industry specialists …