Visual quality inspection in manufacturing, needs good labeled data. Nowadays, external labelers and machine learning engineers are curating those datasets for the manufacturers. It is usually a lengthy and expensive process. We present our approach for autonomous visual inspection systems that let manufacturing operators curate and label visual quality data directly during their work and therefore enable the most accurate visual inspection systems at a reduced cost and with no data labeling required by external parties.

Download the slides for this talk.Download ( PDF, 1249.26 MB)