ITESOFT and the Computer Vision Centre (CVC) develop together a transfer of knowledge program in the framework of the Marie Curie Actions, Industry-Academia Partnerships and Pathways (IAPP).
The driving force of this program is the optimization of an automatic document processing system.
Public organizations, social security services and large companies handle every day a large volume of various administrative documents: identity cards, forms, mails, etc. They look for help in process optimization, cost reduction and customer satisfaction. Automation is the answer.
Administrative documents are textual documents describing a demand or justifying a demand. The clerk who processes the document has to identify the demand (document classification) and extract the few indices relevant for the demand (document reading). The information is very heterogeneous:
- Information can be explicit, directly readable within the document, or implicit, i.e. not directly readable but inferred from readable information.
- Information can be machine printed, hand-printed or handwritten.
- Information can be structured like forms or unstructured like free text in a natural language. Notice that there is a vast amount of different administrative document structures.
- Information can be isolated and easy to locate (printed or written on a white background) or overlapping and hard to segment (printed on a textured background or written over a printed text, …).
- Information can be in color.
The French company ITESOFT has developed products to automate the classification and the data extraction from administrative documents. These products rely on key-technologies: Graphs matching algorithms, Handwritten wordspotting, Document classification, Document segmentation, Color image processing, etc.
The Document Analysis and Pattern Recognition Group of the Computer Vision Center (CVC) in Barcelona has large R&D experience in the field of Document Image Analysis and Recognition. The group is active in research in symbol recognition, indexing and browsing by graphical content, sketchy interfaces, diagrammatic reasoning and visual languages for graphic documents, graphics recognition architectures, reading systems for forms and structured documents, camera-based OCR.
We identified several knowledge areas and skills that can be exchanged between ITESOFT and CVC and combined to breakthrough the current state of the art.
This program mainly involves staff exchanges. We also plan recruitments and workshops with researchers from outside the consortium. The staff exchanges typically lead small collaborative teams to work together on the key-technologies.