Convolutional Dictionary Learning for Multi-Channel Signals

There has recently been a rapid growth in interest in the design of efficient algorithms for convolutional sparse coding, and in the application of these methods to signal and image processing inverse problems. Thus far, however, the design of algorithms and methods for multi-channel signals has received very little attention. In this work we extend our initial results in convolutional sparse coding and dictionary learning for this type of data, proposing new algorithms that scale well to signals with large numbers of channels, and demonstrate their performance in an application involving hyperspectral imagery.