Recognise Official Document

  • Last Post 05 April 2018
m.aquilina posted this 29 March 2018

Is it possible to scan an official document (such as an official driving license) and have the ocr recognise the document type (eg. uk passport, uk driving license, italian passport) together with sorted fields (name, surname etc)?

Order By: Standard | Newest | Votes
Oksana S. posted this 02 April 2018

Thank you for your interest in ABBYY products!

Could you please specify if it is principal to use an online service for OCR and not an offline software? The fact is that for your usage scenario we would recommend to try any of our Data and Document Capture software, for example:

  • ABBYY FlexiCapture – this is our offline ready-to-use solution designed to process the structured document,
  • ABBYY FlexiCapture Engine - it is a non-cloud based data capture SDK, that provides tools for extracting data from documents that have or have no fixed layout.

FlexiCapture and FlexiCapture Engine are intended for data processing and document capturing, when Cloud OCR SDK is mostly used for full-page recognition, although it has API for field-level recognition with some limitations (no analysis is available). Cloud OCR SDK is not specially designed for the identification documents processing and it does not support using of the document definition file as it can be done in our offline Data Capture products. These Data Capture solutions provide tools for extracting data from different documents, even those that have no fixed layout (Cloud OCR SDK can be used only for static forms). If this is what you are looking for, please contact your Sales Manager in your region office.

m.aquilina posted this 02 April 2018

Thanks for replying! We are looking to extract data from official documents such as driving licenses and passports. Since these are standard, we were wondering whether a service on the cloud is already available that does this and not have us do the identification and reinvent the wheel ... is this at all possible? For example, can we somehow register 3 different types of documents with regions and have ABBYY FlexiCapture recognise the right document from an image and provide the respective fields? 

Oksana S. posted this 05 April 2018

Sorry for the delay in response.

In order to use ABBYY FlexiCapture for processing such documents first you will have to create your own definitions of these documents. Now ABBYY does not have ready solutions for recognition of UK or Italian documents. But this can be done by our Professional Services (if you are interested, please contact your region Sales Manager for discussing the price).

Currently there are available the following solutions for official documents processing:

For new documents which have a machine-readable zone (MRZ), you can try the processMRZ method. It finds MRZ on the image and extracts data from it.

In case you would like to recognize documents without MRZ, you can perform full-text recognition and use the processImage method. Also you could export recognized data in XML – this output will contain the text and the text coordinates. In this case you have to implement line-by-line parsing of the recognized text to extract the certain fields.

It is special developer tool which is intended for Russian and CIS passports, ID cards and Driving licenses. Unfortunately, this solution is not applicable to European passports and it’s project modification is only possible by ABBYY Professional Services. This solution is based on ABBYY FlexiCapture technologies, so a developer can make similar application by using ABBYY FlexiCapture Engine.

All listed variants require some developing either by you, or by our Professional Services.