In this thesis we introduce some techniques reaching close to state of the art performance in the topic of spectrum reconstruction while requiring in input only a few HS images or even pixels. We also show the importance of exploiting the physical model in the training pipeline to constraint the output and ease out the whole process.

Semi-supervised deep learning techniques for spectrum reconstruction from RGB images

Simonetto, Adriano
2019/2020

Abstract

In this thesis we introduce some techniques reaching close to state of the art performance in the topic of spectrum reconstruction while requiring in input only a few HS images or even pixels. We also show the importance of exploiting the physical model in the training pipeline to constraint the output and ease out the whole process.
2019-09-10
networks, hyperspectral, reconstruction, learning, images​
File in questo prodotto:
File Dimensione Formato  
thesis_adriano_simonetto.pdf

Open Access dal 11/09/2022

Dimensione 8.66 MB
Formato Adobe PDF
8.66 MB Adobe PDF Visualizza/Apri

The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.12608/28899