Skip to main content

Open Neural Networks Hardware Implementation for Defect Classification Systems

Objective

Real-time classification of visible defects on flat and moving materials is a critical issue in quality control and management in many process-oriented industries.

ONNI-ADC aims to develop an advanced defect classification system prototype which will be based on a neural network system approach and OMI technologies.

The prototype will be tested in a real user manufacturing plant to validate the system performances and the results for the targeted markets.

The ONNI-ADC development process will consist of the identification of the needs and the requirements of potential users in Europe in sectors such as steel, aluminium and paper; the development of OMI neural network supercells; and the prototyping and testing of the final OMI-based defect classification system in a steel-making plant.

Coordinator

Sema Group SA
Address
16-20 Rue Barbes
92126 Montrouge
France

Participants (4)

ANSALDO RICERCHE SPA*
Italy
Address
Corso Perrone 25
16152 Genova
CENTRO SVILUPPO MATERIALI SPA
Italy
Address
Via Di Castel Romano 100/102
Roma
Ilva SpA
Italy
Address
Viale Castro Pretorio 122
00185 Roma
UNIVERSITA DEGLI STUDI DI GENOVA
Italy
Address
Via Opera Pia 11 A
16145 Genova