Active-learning-driven deep interactive segmentation for cost-effective labeling of crop-weed image data

Autor(en): Sikouonmeu, Freddy
Atzmueller, Martin 
Herausgeber: Stein, A.
Hoffmann, C.
Ruckelshausen, A.
Steckel, T.
Helga, F.
Muller, H.
Stichwörter: active learning; Cost effectiveness; Cost reduction; crop-weed detection; Crops; data annotation; deep interactive segmentation; deep learning; Electric loads; Interactive segmentation; Large dataset; Learning systems; Pixels; real-time semantic segmentation; Real-time semantics; Semantic segmentation; Semantics; Supervised learning; Weed control; Weed detection
Erscheinungsdatum: 2023
Herausgeber: Gesellschaft fur Informatik (GI)
Journal: Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI)
Volumen: P-330
Startseite: 507 – 512
Active learning has shown its reliability in (semi-)supervised machine learning tasks to reduce the labeling cost for large datasets in various areas. However, in the agricultural field, despite past attempts to reduce the labeling cost and the burden on the labeler in acquiring image labels, the load during the acquisition of pixel-level labels for semantic image segmentation tasks remains high. Typically, the respective pixel-level masks are acquired manually by drawing polygons over irregular and complex-shaped object boundaries. In contrast, this paper proposes a method leveraging the power of a click-based deep interactive segmentation model (DISEG) in an active learning approach to harvest high-quality image segmentation labels at a low cost for training a real-time task model by only clicking on the objects' fore- and background surfaces. Our first experimental results indicate that with an average of 3 clicks per image object and using only 3% of the unlabeled dataset, we can acquire pixel-level labels with good quality at low cost. © 2023 Gesellschaft fur Informatik (GI). All rights reserved.
Cited by: 0; Conference name: 43. Jahrestagung der Gesellschaft fur Informatik in der Land-, Forst- und Ernahrungswirtschaft - Resiliente Agri-Food-Systeme: Herausforderungen und Losungsansatze, GIL 2023 - 43rd Annual Conference of the Society for Informatics in Agriculture, Forestry, and Food Industry - Resilient Agri-Food Systems: Challenges and Solutions, GIL 2023; Conference date: 13 February 2023 through 14 February 2023; Conference code: 193633
ISBN: 9783885797241
ISSN: 1617-5468
Externe URL:

Show full item record

Page view(s)

Last Week
Last month
checked on May 24, 2024

Google ScholarTM