Liu, Chenying und Albrecht, Conrad M und Wang, Yi und Zhu, Xiao Xiang (2024) CromSS: Cross-Modal Pre-Training with Noisy Labels for Remote Sensing Image Segmentation. In: International Conference on Learning Representations, ICLR, Seiten 1-7. ICLR 2024, 2024-05-07, Vienna, Austria. doi: 10.48550/arXiv.2405.01217.
![]() |
PDF
302kB |
Offizielle URL: https://ml-for-rs.github.io/iclr2024
Kurzfassung
We study the potential of noisy labels y to pretrain semantic segmentation models in a multi-modal learning framework for geospatial applications. Specifically, we propose a novel Cross-modal Sample Selection method (CromSS) that utilizes the class distributions $P^{(d)}(x,c)$ over pixels $x$ and classes c modelled by multiple sensors/modalities $d$ of a given geospatial scene. Consistency of predictions across sensors d is jointly informed by the entropy of $P^{(d)}(x,c)$. Noisy label sampling we determine by the confidence of each sensor d in the noisy class label, $P^{(d)}(x,c=y(x))$. To verify the performance of our approach, we conduct experiments with Sentinel-1 (radar) and Sentinel-2 (optical) satellite imagery from the globally-sampled SSL4EO-S12 dataset. We pair those scenes with 9-class noisy labels sourced from the Google Dynamic World project for pretraining. Transfer learning evaluations (downstream task) on the DFC2020 dataset confirm the effectiveness of the proposed method for remote sensing image segmentation.
elib-URL des Eintrags: | https://elib.dlr.de/204344/ | ||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Dokumentart: | Konferenzbeitrag (Vortrag) | ||||||||||||||||||||
Titel: | CromSS: Cross-Modal Pre-Training with Noisy Labels for Remote Sensing Image Segmentation | ||||||||||||||||||||
Autoren: |
| ||||||||||||||||||||
Datum: | 2024 | ||||||||||||||||||||
Erschienen in: | International Conference on Learning Representations, ICLR | ||||||||||||||||||||
Referierte Publikation: | Ja | ||||||||||||||||||||
Open Access: | Ja | ||||||||||||||||||||
Gold Open Access: | Nein | ||||||||||||||||||||
In SCOPUS: | Nein | ||||||||||||||||||||
In ISI Web of Science: | Nein | ||||||||||||||||||||
DOI: | 10.48550/arXiv.2405.01217 | ||||||||||||||||||||
Seitenbereich: | Seiten 1-7 | ||||||||||||||||||||
Status: | veröffentlicht | ||||||||||||||||||||
Stichwörter: | pre-training, noisy labels, semantic segmentation, multi-modal deep learning, sample selection, SSL4EO-S12 dataset, geospatial artificial intelligence | ||||||||||||||||||||
Veranstaltungstitel: | ICLR 2024 | ||||||||||||||||||||
Veranstaltungsort: | Vienna, Austria | ||||||||||||||||||||
Veranstaltungsart: | internationale Konferenz | ||||||||||||||||||||
Veranstaltungsdatum: | 7 Mai 2024 | ||||||||||||||||||||
HGF - Forschungsbereich: | Luftfahrt, Raumfahrt und Verkehr | ||||||||||||||||||||
HGF - Programm: | Raumfahrt | ||||||||||||||||||||
HGF - Programmthema: | Erdbeobachtung | ||||||||||||||||||||
DLR - Schwerpunkt: | Raumfahrt | ||||||||||||||||||||
DLR - Forschungsgebiet: | R EO - Erdbeobachtung | ||||||||||||||||||||
DLR - Teilgebiet (Projekt, Vorhaben): | R - Künstliche Intelligenz, R - Optische Fernerkundung | ||||||||||||||||||||
Standort: | Oberpfaffenhofen | ||||||||||||||||||||
Institute & Einrichtungen: | Institut für Methodik der Fernerkundung > EO Data Science | ||||||||||||||||||||
Hinterlegt von: | Albrecht, Conrad M | ||||||||||||||||||||
Hinterlegt am: | 27 Mai 2024 09:53 | ||||||||||||||||||||
Letzte Änderung: | 29 Mai 2024 15:37 |
Nur für Mitarbeiter des Archivs: Kontrollseite des Eintrags