IDscan Semantic Engine
Project overview:
IDscan Semantic Engine is a software library used in some of IDscan products. The system can recognise and semantically extract “useful” information from scanned images of any document type. What is unique about IDscan Semantic Engine is the capability of recognising any document in any format, the document need not to be structured nor have fixed layout. Semantic engine uses advanced OCR technology and Natural Language Processing to understand the document content and classify/extract useful pieces of information.
Semantic Engine is currently being used by many banks in the world to perform their KYC processes and process large sets of legacy scanned documents. It has proved its efficiency and uniqueness in automatically extracting full names, addresses, currencies, and tables.
Semantic Engine was built using C#, C++ programming languages, and advanced image processing, natural language processing algorithms.
Project team
Zaher Joukhadar, Moneer Allito, Anas Khayata, Osama Makansi, Raniem Arour, Baraa Abo Helal, and Meltem Cetiner
Technologies:
My role in the project:
I helped in conceiving, exploring, and elaborating the project idea. I led the initial effort to build the system, and then was involved intermittently in developing and leading the effort in the project.
Patents & Awards:
Several patents applications are in progress, I will list them when they are published.
The project received a financial award from the Turkish government.