Filtered Residual Compression for Satellite Images - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Filtered Residual Compression for Satellite Images

Résumé

Learned image compression neural networks have difficulties adapting to certain satellite image characteristics, especially high frequencies that disappear at a high bit-rate in the blur generated in the reconstruction. To answer this problem we describe a joint end-to-end trainable neural network. It is separated into a general compression network and a smaller specialised network. We train a specialized network to compress the residual part of the image to best preserve the highfrequency details present in the satellite images. The proposed model achieves higher rate-distortion performance than current lossy image compression standards and also manages to retrieve details previously poorly reconstructed.
Fichier principal
Vignette du fichier
IGARSS_2023_BACCHUS.pdf (426.25 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04125811 , version 1 (12-06-2023)

Licence

Paternité

Identifiants

  • HAL Id : hal-04125811 , version 1

Citer

Pascal Bacchus, Renaud Fraisse, Christine Guillemot, Aline Roumy. Filtered Residual Compression for Satellite Images. IGARSS 2023 - International Geoscience and Remote Sensing Symposium, IEEE Geoscience and Remote Sensing Society, Jul 2023, Pasadena, CA, United States. pp.1-1. ⟨hal-04125811⟩
46 Consultations
43 Téléchargements

Partager

Gmail Facebook X LinkedIn More