This project is targeted at the whole sector of manufacturing, assembly and production. Almost any company that makes products using an automated production line needs some kind of quality control. Very often their quality control includes visual inspection. If this inspection task is to be automated in a machine vision system, one has to solve the problem of how to implement a human decision-making process (good part vs. bad part) in software. Currently, this requires a step-by-step reprogramming or parametrization of the software, which may last for several months until satisfying results are obtained. The results of DYNAVIS will enable us to use human-machine cooperation to learn complicated inspection tasks instead of step-by-step improvements and adaptations of software.

In order to create such methods we will focus on the following scientific objectives:

• machine learning methods for processing the complicated data produced by the vision system.
• methods to deal with multiple, possibly contradictory teaching input by the operators.
• methods for predicting success or failure of the learning process in early stages of the training process.

In all of these topics we will concentrate on making machine vision robust and easily applicable in industrial environments.