Convolutional sparse representations have recently attracted significant attention from the imaging community, thanks to their structural properties and their success in numerous imaging applications, ranging from restoration to super resolution and HDR imaging. Despite the development of efficient sparse coding and dictionary learning algorithms, several theoretical as well as practical aspects of these representations are still not thoroughly understood, and convolutional sparse representations remain, as a technique, far less mature than standard sparse representations. This minisymposium provides a selection of recent progress in this area, including dictionary learning algorithms, connections with deep learning, and new applications.

Online Convolutional Dictionary Learning for Multimodal Imaging

Ulugbek Kamilov (Washington University in St. Louis)

From convolutional analysis operator learning (CAOL) to convolutional neural network (CNN)

Il Yong Chun (University of Michigan)

Greedy and learned approaches for convolutional sparse coding

Raja Giryes (Tel Aviv University)

Convolutional Sparse Coding vs Aggregation of Independent Estimates

Diego Carrera (Politecnico di Milano)

Organizers:

Giacomo Boracchi (Politecnico di Milano)

Alessandro Foi (Tampere University of Technology)

Brendt Wohlberg (Los Alamos National Laboratory)

Keywords:

convolutional sparsity, deep learning, image enhancement, image reconstruction, image representation, machine learning