Compressive Sensing via Convolutional Factor Analysis

We solve the compressive sensing problem via the convolutional dictionary learning. We develop an alternating direction method of multipliers (ADMM) paradigm for compressive inversion. The proposed algorithm can provide the reconstructed images as well as the features, which can be directly used for classification or clustering tasks. A joint compressive sensing inversion and classification regime is presented. Reconstruction results and classification results on benchmark datasets provided.