Transformers in Remote Sensing
Transformer-based architectures have recently achieved remarkable success across machine learning domains due to their ability to model complex dependencies using attention mechanisms. In remote sensing—where data is typically high-dimensional, multimodal, and temporally rich—transformers offer a compelling framework for integrated analysis and interpretation.
This special session aims to bring together recent advances in applying transformer models to remote sensing data, with a focus on multimodal learning, data fusion, and domain-specific adaptations. Contributions are encouraged from both theoretical and applied perspectives, addressing challenges in Earth observation using attention-based architectures.
The sessions covers (but is not limited to) papers on
- Vision Transformers (ViT) for satellite and aerial image interpretation
- Spatio-temporal transformers for change detection and time series modeling
- Cross-modal attention for fusion of optical, SAR, LiDAR, and hyperspectral data
- Multimodal transformers integrating sensor metadata with imagery
- Self-supervised learning and pretraining strategies in geospatial contexts
- Semantic segmentation, classification, and object detection in remote sensing
- Lightweight transformer models for onboard or edge deployment
- Real-world applications: agriculture, land cover mapping, disaster response, environmental monitoring
Chairman: Dr. Erdem Akagündüz
Graduate School of Informatics, Middle East Technical University (METU), Türkiye
Erdem Akagündüz is currently an Associate Professor with the Graduate School of Informatics at Middle East Technical University (METU), Türkiye. He is the principal investigator of the Applied Intelligence Research Laboratory (AIRLab), based at METU Informatics Institute. He is professionally interested in computer vision, deep learning, pattern recognition, image processing, machine learning, object tracking, and 3D modelling. He has authored numerous journal and conference papers and holds several international patents in these areas. In addition to his academic research, he provides consultancy to industry partners on AI-based technology development. He received his Ph.D. from METU and has held research positions at the University of York and ASELSAN Inc.