## Resumen

Given a set of dictionary filters, the most widely used formulation of the convolutional sparse coding (CSC) problem is Convolutional BPDN (CBPDN), in which an image is represented as a sum over a set of convolutions of coefficient maps; usually, the coefficient maps are ℓ_{1}-norm penalized in order to enforce a sparse solution. Recent theoretical results, have provided meaningful guarantees for the success of popular ℓ_{1}-norm penalized CSC algorithms in the noiseless case. However, experimental results related to the ℓ_{0}-norm penalized CSC case have not been addressed.In this paper we propose a two-step ℓ_{0}-norm penalized CSC (ℓ_{0}-CSC) algorithm, which outperforms (convergence rate, reconstruction performance and sparsity) known solutions to the ℓ_{0}-CSC problem. Furthermore, our proposed algorithm, which is a convolutional extension of our previous work [1], originally develop for the ℓ_{0} regularized optimization problem, includes an escape strategy to avoid being trapped in a saddle points or in inferior local solutions, which are common in nonconvex optimization problems, such those that use the ℓ_{0}-norm as the penalty function.

Idioma original | Inglés |
---|---|

Título de la publicación alojada | Proceedings of the 2018 IEEE 25th International Conference on Electronics, Electrical Engineering and Computing, INTERCON 2018 |

Editorial | Institute of Electrical and Electronics Engineers Inc. |

ISBN (versión digital) | 9781538654903 |

DOI | |

Estado | Publicada - 6 nov. 2018 |

Evento | 25th IEEE International Conference on Electronics, Electrical Engineering and Computing, INTERCON 2018 - Lima, Perú Duración: 8 ago. 2018 → 10 ago. 2018 |

### Serie de la publicación

Nombre | Proceedings of the 2018 IEEE 25th International Conference on Electronics, Electrical Engineering and Computing, INTERCON 2018 |
---|

### Conferencia

Conferencia | 25th IEEE International Conference on Electronics, Electrical Engineering and Computing, INTERCON 2018 |
---|---|

País/Territorio | Perú |

Ciudad | Lima |

Período | 8/08/18 → 10/08/18 |

## Huella

Profundice en los temas de investigación de 'Fast Convolutional Sparse Coding with ℓ_{0}Penalty'. En conjunto forman una huella única.