Kapitel/Beitrag aus dem Buch: Heizmann M. & Längle T. 2020. Forum Bildverarbeitung 2020.
One way for the visual inspection of assemblies with
many variants is to compare camera images with the corresponding
rendered view of the CAD model. In this paper, we
address the problem to decide whether there are significant
differences between camera and rendered images, which signal
an assembly error. Our approach uses a Conditional Generative
Adversarial Network (CGAN) to translate the camera
image to a rendered like one, followed by error detection by
comparing the translated and rendered images.
Datha N. & Thiel M. 2020. Measuring similarity of rendered and realimage pairs using domain translation byemploying Conditional GenerativeAdversarial Networks. In: Heizmann M. & Längle T (eds.), Forum Bildverarbeitung 2020. Karlsruhe: KIT Scientific Publishing. DOI: https://doi.org/10.58895/ksp/1000124383-22
This chapter distributed under the terms of the Creative Commons Attribution + ShareAlike 4.0 license. Copyright is retained by the author(s)
Dieses Buch ist Peer reviewed. Informationen dazu Hier finden Sie mehr Informationen zur wissenschaftlichen Qualitätssicherung der MAP-Publikationen.
Veröffentlicht am 25. November 2020