Hyper-parameter selection on convolutional dictionary learning through local `0,∞ norm

Gustavo Silva, Jorge Quesada, Paul Rodriguez

Producción científica: Capítulo del libro/informe/acta de congresoContribución a la conferenciarevisión exhaustiva


Convolutional dictionary learning (CDL) is a widely used technique in many applications on the signal/image processing and computer vision fields. While many algorithms have been proposed in order to improve the computational run-time performance during the training process, a thorough analysis regarding the direct relationship between the reconstruction performance and the dictionary features (hyper-parameters), such as the filter size and filter bank's cardinality, has not yet been presented. As arbitrarily configured dictionaries do not necessarily guarantee the best possible results during the test process, a correct selection of the hyper-parameters would be very favorable in the training and testing stages. In this context, this works aims to provide an empirical support for the choice of hyper-parameters when learning convolutional dictionaries. We perform a careful analysis of the effect of varying the dictionary's hyper-parameters through a denoising task. Furthermore, we employ a recently proposed local `0,∞ norm as a sparsity measure in order to explore possible correlations between the sparsity induced by the learned filter bank and the reconstruction quality at test stage.

Idioma originalInglés
Título de la publicación alojadaEUSIPCO 2019 - 27th European Signal Processing Conference
EditorialEuropean Signal Processing Conference, EUSIPCO
ISBN (versión digital)9789082797039
EstadoPublicada - set. 2019
Evento27th European Signal Processing Conference, EUSIPCO 2019 - A Coruna, Espana
Duración: 2 set. 20196 set. 2019

Serie de la publicación

NombreEuropean Signal Processing Conference
ISSN (versión impresa)2219-5491


Conferencia27th European Signal Processing Conference, EUSIPCO 2019
CiudadA Coruna


Profundice en los temas de investigación de 'Hyper-parameter selection on convolutional dictionary learning through local `0,∞ norm'. En conjunto forman una huella única.

Citar esto