Adaptive Rank Estimation Based Tensor Factorization Algorithm for Low-Rank Tensor Completion

As the generalized form of vectors and matrices, tensors can describe the data structures more directly. In the process of acquiring higher-order tensor, some entries may be lost due to various reasons. Low-rank tensor completion (LRTC) is to recover the missing entries according to the low-rank pro...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Chinese Control Conference s. 3444 - 3449
Hlavní autori: Liu, Han, Liu, Jing, Su, Liyu
Médium: Konferenčný príspevok..
Jazyk:English
Vydavateľské údaje: Technical Committee on Control Theory, Chinese Association of Automation 01.07.2019
Predmet:
adaptive rank estimation
Data structures
Estimation
image completion
Iterative algorithms
Low-rank tensor completion
Tensors
tubal rank
Vectors
ISSN:1934-1768
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:As the generalized form of vectors and matrices, tensors can describe the data structures more directly. In the process of acquiring higher-order tensor, some entries may be lost due to various reasons. Low-rank tensor completion (LRTC) is to recover the missing entries according to the low-rank property of tensor. The tensor completion by tensor factorization (TCTF) method was proposed to solve the LRTC problem. However, tubal rank of the original tensor is assumed known in TCTF method. In this paper, we propose an adaptive rank estimation based tensor factorization (ARE-TF) algorithm for low-rank tensor completion. Based on Ll-norm regularized rank-one matrix completion algorithm, the estimated rank is obtained by iterative updating procedures. The experiment results in image completion show that the proposed algorithm has better performance than the existing algorithms.
ISSN:1934-1768
DOI:10.23919/ChiCC.2019.8865482