Benutzer: Gast  Login
Titel:

Towards Fully-Synthetic Training for Industrial Applications

Dokumenttyp:
Konferenzbeitrag
Autor(en):
Mayershofer, C.; Ge, T.; Fottner, J.
Nicht-TUM Koautoren:
nein
Kooperation:
-
Abstract:
This paper proposes a scalable approach for synthetic image generation of industrial objects leveraging Blender for image rendering. In addition to common components in synthetic image generation research, three novel features are presented: First, we model relations between target objects and randomly apply those during scene generation (Object Relation Modelling (ORM)). Second, we extend the idea of distractors and create Object-alike Distractors (OAD), resembling the textural appearance (i.e....     »
Stichworte:
Object detection, Synthetic data, Domain randomization
Intellectual Contribution:
Contribution to Practice
Kongress- / Buchtitel:
10th International Conference on Logistics, Informatics and Service Sciences (LISS)
Kongress / Zusatzinformationen:
Budapest, Hungary
Jahr:
2020
Volltext / DOI:
doi:no DOI available
Key publication:
Ja
Peer reviewed:
Ja
International:
Ja
commissioned:
not commissioned
Interdisziplinarität:
Ja
Leitbild:
;
Technology:
Ja
Ethics und Sustainability:
Nein
 BibTeX
Versionen