We propose an automatic parameter selection strategy for variational image super-resolution of blurred and down-sampled images corrupted by additive white Gaussian noise (AWGN) with unknown standard deviation. By exploiting particular properties of the operators describing the problem in the frequency domain, our strategy selects the optimal parameter as the one optimising a suitable residual whiteness measure. Numerical tests show the effectiveness of the proposed strategy for generalised ℓ2 - ℓ2 Tikhonov problems.
Residual Whiteness Principle for Automatic Parameter Selection in ℓ2 - ℓ2 Image Super-Resolution Problems / Pragliola M.; Calatroni L.; Lanza A.; Sgallari F.. - STAMPA. - 12679:(2021), pp. 476-488. (Intervento presentato al convegno 8th International Conference on Scale Space and Variational Methods in Computer Vision, SSVM 2021 tenutosi a Cabourg, France (online) nel 2021) [10.1007/978-3-030-75549-2_38].
Residual Whiteness Principle for Automatic Parameter Selection in ℓ2 - ℓ2 Image Super-Resolution Problems
Pragliola M.
Primo
Membro del Collaboration Group
;Lanza A.Penultimo
Membro del Collaboration Group
;Sgallari F.Ultimo
Membro del Collaboration Group
2021
Abstract
We propose an automatic parameter selection strategy for variational image super-resolution of blurred and down-sampled images corrupted by additive white Gaussian noise (AWGN) with unknown standard deviation. By exploiting particular properties of the operators describing the problem in the frequency domain, our strategy selects the optimal parameter as the one optimising a suitable residual whiteness measure. Numerical tests show the effectiveness of the proposed strategy for generalised ℓ2 - ℓ2 Tikhonov problems.File | Dimensione | Formato | |
---|---|---|---|
Residual Whiteness Principle for Automatic Parameter Selection in ℓ2 - ℓ2 Image Super-Resolution Problems.pdf
accesso aperto
Tipo:
Postprint
Licenza:
Licenza per accesso libero gratuito
Dimensione
2.24 MB
Formato
Adobe PDF
|
2.24 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.