Open AccessTri‐decomposition model for image recovery
Author(s) -
Jin Mengying,
Chen Yunjie,
Zhang Fanlong
Publication year - 2018
Publication title -
electronics letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.375
H-Index - 146
eISSN - 1350-911X
pISSN - 0013-5194
DOI - 10.1049/el.2018.5718
Subject(s) - rank (graph theory) , noise (video) , sparse matrix , image (mathematics) , computer science , component (thermodynamics) , sparse approximation , matrix (chemical analysis) , artificial intelligence , pattern recognition (psychology) , decomposition , image restoration , face (sociological concept) , matrix decomposition , computer vision , image processing , mathematics , ecology , social science , eigenvalues and eigenvectors , physics , materials science , composite material , combinatorics , quantum mechanics , sociology , biology , gaussian , thermodynamics
This Letter reveals the feasibility of decomposing a matrix into three component matrices for image recovery. Image recovery emerges in many areas, such as image processing, computer vision, and pattern recognition. Recently, the low rank assumption‐based image recovery methods catch the researcher's attention. The authors assume the real data matrix has low rank and the error matrix is sparse. However, they are limited to the low‐rank component being exactly low‐rank, and the sparse component being exactly sparse. Either or both these assumptions are not exactly satisfied in practice and should be relaxed. This Letter presents a tri‐decomposition method for dealing with the image data corrupted by both large sparse noise and small dense noise. The method parts the observed data into the clean data, sparse noise, and dense noise by different measure functions. Extensive experiments on face images and surveillance videos demonstrate the effectiveness of the proposed method.