Enrico Marchi

Enrico Marchi

Ph.D XXXVIII

Supervisor: Gian Luca Foresti

Phone: 

Room: 

Mail: marchi.enrico001@spes.uniud.it

Research Project

Anomaly detection for 3D objects in industrial settings

Quality control in industry has been done traditionally by humans which implies high probabilities at committing mistakes because the brain is not able to keep the attention at its best for long time. Machines are the best option to fulfil this task, however it is not as easy as it seams to accomplish it. Indeed, the geometry of the items varies a lot and mostly the data about faulty objects are scarce. The first thing to do is to understand how data should be physically taken from 3D objects. Then another key step of this research is to augment the data generating synthetic ones which helps to collect data about wrong items. Finally Deep Learning and other Machine Vision techniques should be investigated to find the ones that best fit the problem.