We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Quantifizierung der Klassifikationsleistung von Oberflächeninspektionssystemen in der Flachstahlproduktion.
- Authors
Brandenburger, Jens; Krippendorff, Fabian; Krätzner, Michael; Nörtersheuser, Michael; Chen, Xin; Boss, Andreas; Jonker, Kees; Pipard, Nicolas; Ebner, Andreas
- Abstract
In modern steel production, automatic surface inspection systems (ASIS) for the detection and classification of surface defects are widely used and their results have gained strongly in importance. Nevertheless, there is still a lack of recognized methods for an objective and comprehensive performance evaluation of the systems able to determine suitable parameters for the ASIS classification assessment in real production operations with reasonable effort. This paper addresses the problem of estimating the recall in the presence of unknown "ground truth", as a key measure for the capability evaluation of classifying image processing systems. A methodology for recall estimation using an auxiliary classifier is presented and research needs for its practical implementation are identified.
- Subjects
IMAGE processing; SURFACE defects; IMAGING systems; STEEL; CLASSIFICATION
- Publication
Technisches Messen, 2023, Vol 90, Issue 10, p650
- ISSN
0171-8096
- Publication type
Article
- DOI
10.1515/teme-2023-0035