Mobi 3D: Recognition of 3D objects
MOBI 3D (Mobile Object a Bridge to Immersive 3D): 2015-2016 collaborative research project labelled by the "pôle images et réseaux" and financed by the Région Bretagne, Brest Métropole and the Conseil général du Finistère.
Consortium of 3 partners including 1 academic: bookBeo (SME project leader), E-Mage-In 3D (SME) and IMT - Lab TOMS (academic).
The objective of the research was to recognize reality and to associate it with 3D content ready to be printed in any type of material: metal, resin, algae.
The consortium designed and tested algorithms for visual recognition by 2D and 3D capture, machine learning and deep learning.
Two research methods and their application solutions for industry were tested: deep learning and 3D capture (3D descriptors).
The result of the research has enabled us to advance the state of the art on the recognition of real and 3D objects and to offer cognitive assistance tools to industrial operators.
The redo-tracking solution for assisting operators in controlling defects is the result of this research.