ImageReCon

An (mn) by (M-m+1)(N-n+1) matrix.

mm, nn

Size of the small patches (m by n).

m, n

(optional) The stepsize when splicting the image. Default is 1

stepsize

Suppose an image matrix A of size M by N is splited into (M-m+1)(N-n+1)
overlapped small patches of size m by n. Given a matrix B of size (mn) by (M-m+1)(N-n+1)
, where each column of B is a vectorized small patch, this function try to
reconstruct the full image A.

Method proposed in "Simultaneous Sparse Dictionary Learning and Pruning" ( Qu and Wang (2016) <arXiv:1605.07870>) is implemented. The idea is to conduct a linear decomposition of a signal using a few atoms of a learned and usually over-completed dictionary instead of a pre-defined basis. A proper size of the to-be-learned dictionary is determining at the same time during the procedure. Application includes image denoising and image inpainting.

ImageReCon: Combine patches of image into full image

Description

Usage

Arguments

Value