EO-VAE: Towards A Multi-sensor Tokenizer for Earth Observation Data
This addresses the problem of efficient latent representation for Earth observation data, which is incremental as it adapts existing tokenizer paradigms to a specific domain.
The paper tackles the challenge of tokenizing Earth observation data with diverse sensors by proposing EO-VAE, a multi-sensor variational autoencoder that uses dynamic hypernetworks to handle flexible channel combinations, achieving superior reconstruction fidelity compared to TerraMind tokenizers on the TerraMesh dataset.
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.