Online Convolutional Dictionary Learning for Multimodal Imaging

TL;DR

This paper introduces an online convolutional dictionary learning method for multimodal imaging that leverages redundancies across modalities, combining group-sparse representations with TV regularization for improved image reconstruction.

Contribution

It presents a novel online algorithm for unsupervised learning of convolutional dictionaries tailored for large-scale multimodal imaging datasets.

Findings

Enhanced image reconstruction quality in joint intensity-depth imaging.

Effective unsupervised learning of convolutional dictionaries on large datasets.

Combines group-sparse and TV regularization for multimodal image enhancement.

Abstract

Computational imaging methods that can exploit multiple modalities have the potential to enhance the capabilities of traditional sensing systems. In this paper, we propose a new method that reconstructs multimodal images from their linear measurements by exploiting redundancies across different modalities. Our method combines a convolutional group-sparse representation of images with total variation (TV) regularization for high-quality multimodal imaging. We develop an online algorithm that enables the unsupervised learning of convolutional dictionaries on large-scale datasets that are typical in such applications. We illustrate the benefit of our approach in the context of joint intensity-depth imaging.