We present an efficient method for the reconstruction of multispectral information from RGB images, as part of the NTIRE 2022 Spectral Reconstruction Challenge. Given an input image, our method determines a global RGB-to-spectral linear transformation matrix, based on a search through optimal matrices from training images that share low-level features with the input. The resulting spectral signatures are then adjusted by a global scaling factor, determined through a lightweight SqueezeNet-inspired neural network. By combining the efficiency of linear transformation matrices with the data-driven effectiveness of convolutional neural networks, we are able to achieve superior performance than winners of the previous editions of the challenge.

Agarla, M., Bianco, S., Buzzelli, M., Celona, L., Schettini, R. (2022). Fast-n-Squeeze: towards real-time spectral reconstruction from RGB images. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2022 (pp.1131-1138) [10.1109/CVPRW56347.2022.00122].

Fast-n-Squeeze: towards real-time spectral reconstruction from RGB images

Agarla, Mirko;Bianco, Simone;Buzzelli, Marco;Celona, Luigi;Schettini, Raimondo
2022

Abstract

We present an efficient method for the reconstruction of multispectral information from RGB images, as part of the NTIRE 2022 Spectral Reconstruction Challenge. Given an input image, our method determines a global RGB-to-spectral linear transformation matrix, based on a search through optimal matrices from training images that share low-level features with the input. The resulting spectral signatures are then adjusted by a global scaling factor, determined through a lightweight SqueezeNet-inspired neural network. By combining the efficiency of linear transformation matrices with the data-driven effectiveness of convolutional neural networks, we are able to achieve superior performance than winners of the previous editions of the challenge.
No
paper
Scientifica
spectral reconstruction
English
2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) - 19-20 June 2022
978-1-6654-8739-9
https://openaccess.thecvf.com/content/CVPR2022W/NTIRE/html/Agarla_Fast-N-Squeeze_Towards_Real-Time_Spectral_Reconstruction_From_RGB_Images_CVPRW_2022_paper.html
Agarla, M., Bianco, S., Buzzelli, M., Celona, L., Schettini, R. (2022). Fast-n-Squeeze: towards real-time spectral reconstruction from RGB images. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2022 (pp.1131-1138) [10.1109/CVPRW56347.2022.00122].
Agarla, M; Bianco, S; Buzzelli, M; Celona, L; Schettini, R
File in questo prodotto:
File Dimensione Formato  
Agarla-2022-CVPR-AAM.pdf

accesso aperto

Descrizione: Proceedings
Tipologia di allegato: Author’s Accepted Manuscript, AAM (Post-print)
Dimensione 6.04 MB
Formato Adobe PDF
6.04 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/10281/392154
Citazioni
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
Social impact