FR2.R10.3
SPATIAL POSITIONAL ENCODING IN ATTENTION-BASED MODELS FOR EARTH OBSERVATION REPRESENTATION LEARNING
Julien Prissimitzis, Julien Michel, Jordi Inglada, CESBIO, France
Session:
FR2.R10: Multimodal Foundation Models for Earth Observation: Unifying Multi-source Geospatial Data for a Holistic Planet Understanding (2/2) Oral
Track:
Community Contributed Themes
Location:
TBD
Presentation Time:
Fri, 14 Aug, 11:30 - 11:45
Session Co-Chairs:
Felix Von Hagen, University of the Bundeswehr Munich and Jiaqi Yang,
Presentation
Discussion
Resources
No resources available.
Session FR2.R10
FR2.R10.1: GEO-MTL: A MULTI-TASK FRAMEWORK FOR JOINT SEMANTIC SEGMENTATION AND TEXTUAL SCENE REPORTING IN REMOTE SENSING
Tadie B. Medimem, Farid Melgani, Sandro Luigi Fiore, University of Trento, Italy; Valentine G. Anantharaj, Oak Ridge National Laboratory, United States
FR2.R10.2: JMC-S1/2: JOINT MASKED AUTOENCODING AND CONTRASTIVE LEARNING FOR MULTIMODAL SENTINEL-1/2 IMAGERY
Chen Qian, University of Chinese Academy of Sciences, China; Tang Liu, Institute of Geographic Sciences and Natural Resources Research, China; Bingxian Wu, University of Chinese Academy of Sciences, China; Yuxiang Lu, China University of Geosciences (Beijing), China; Ling Yao, Chenghu Zhou, Institute of Geographic Sciences and Natural Resources Research, China
FR2.R10.3: SPATIAL POSITIONAL ENCODING IN ATTENTION-BASED MODELS FOR EARTH OBSERVATION REPRESENTATION LEARNING
Julien Prissimitzis, Julien Michel, Jordi Inglada, CESBIO, France
Contacts