Robust generalized low rank approximations of matrices for video denoising

Low rank matrix algorithm has attracted widely attention since it was put forward. There are many algorithms proposed to improve it. Most methods take each sample as a column, but Robust Generalized Low Rank Approximations of Matrices(RGLRAM) treats each sample as a matrix, thus we can find the low...

Full description

Saved in:
Bibliographic Details
Main Authors: Xian Zhao, Gaoyun An, Yigang Cen, Hengyou Wang, Ruizhen Zhao
Format: Conference Proceeding
Language:English
Subjects:
Approximation algorithms
block
Gaussian noise
Image reconstruction
Matrix decomposition
Noise reduction
rank
Robust Generalized Low Rank Approximations of Matrices(RGLRAM)
Robustness
Sparse matrices
video denoising
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Low rank matrix algorithm has attracted widely attention since it was put forward. There are many algorithms proposed to improve it. Most methods take each sample as a column, but Robust Generalized Low Rank Approximations of Matrices(RGLRAM) treats each sample as a matrix, thus we can find the low rank approximations on a collection of matrices not just a single matrix. RGLRAM are not only robust to small Gaussian noise but also to large sparse noise. In these cases, we extend it to video approximation. The video can be treated as a collection of images. Each frame of the video is divided into blocks with equal size according to their coordinates. After the operation of each group, the reconstructed video could be got via splicing the blocks based on these coordinates. Experimental results show that method proposed in this paper is effective.
ISSN:2164-5221
DOI:10.1109/ICSP.2016.7877944