Efficient and Parallel Separable Dictionary Learning
This work addresses the need for efficient dictionary learning in image processing, though it appears incremental by improving computational efficiency over existing methods.
The paper tackled the problem of learning separable dictionaries for 2D signals like images by developing a highly parallelizable algorithm that achieves sparse representations competitive with state-of-the-art methods but at lower computational cost, as demonstrated in image and hyperspectral data representation and image denoising.
Separable, or Kronecker product, dictionaries provide natural decompositions for 2D signals, such as images. In this paper, we describe a highly parallelizable algorithm that learns such dictionaries which reaches sparse representations competitive with the previous state of the art dictionary learning algorithms from the literature but at a lower computational cost. We highlight the performance of the proposed method to sparsely represent image and hyperspectral data, and for image denoising.