EO-VAE: Towards A Multi-sensor Tokenizer for Earth Observation Data
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866914364098871296 |
|---|---|
| author | Lehmann, Nils Wang, Yi Xiong, Zhitong Zhu, Xiaoxiang |
| author_facet | Lehmann, Nils Wang, Yi Xiong, Zhitong Zhu, Xiaoxiang |
| contents | State-of-the-art generative image and video models rely heavily on tokenizers that compress high-dimensional inputs into more efficient latent representations. While this paradigm has revolutionized RGB generation, Earth observation (EO) data presents unique challenges due to diverse sensor specifications and variable spectral channels. We propose EO-VAE, a multi-sensor variational autoencoder designed to serve as a foundational tokenizer for the EO domain. Unlike prior approaches that train separate tokenizers for each modality, EO-VAE utilizes a single model to encode and reconstruct flexible channel combinations via dynamic hypernetworks. Our experiments on the TerraMesh dataset demonstrate that EO-VAE achieves superior reconstruction fidelity compared to the TerraMind tokenizers, establishing a robust baseline for latent generative modeling in remote sensing. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_12177 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | EO-VAE: Towards A Multi-sensor Tokenizer for Earth Observation Data Lehmann, Nils Wang, Yi Xiong, Zhitong Zhu, Xiaoxiang Computer Vision and Pattern Recognition State-of-the-art generative image and video models rely heavily on tokenizers that compress high-dimensional inputs into more efficient latent representations. While this paradigm has revolutionized RGB generation, Earth observation (EO) data presents unique challenges due to diverse sensor specifications and variable spectral channels. We propose EO-VAE, a multi-sensor variational autoencoder designed to serve as a foundational tokenizer for the EO domain. Unlike prior approaches that train separate tokenizers for each modality, EO-VAE utilizes a single model to encode and reconstruct flexible channel combinations via dynamic hypernetworks. Our experiments on the TerraMesh dataset demonstrate that EO-VAE achieves superior reconstruction fidelity compared to the TerraMind tokenizers, establishing a robust baseline for latent generative modeling in remote sensing. |
| title | EO-VAE: Towards A Multi-sensor Tokenizer for Earth Observation Data |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2602.12177 |