Hierarchical Prior Regularized Matrix Factorization for Image Completion

TL;DR

This paper introduces a hierarchical prior regularized matrix factorization model that effectively combines low-rank, total variation, and sparse priors for improved tensor completion, addressing local pattern exploitation.

Contribution

It presents a novel hierarchical regularization approach integrating multiple priors into matrix factorization for tensor completion, enhancing performance over existing methods.

Findings

Outperforms state-of-the-art tensor completion methods

Effectively captures both global and local tensor features

Validated on various datasets with superior results

Abstract

The recent low-rank prior based models solve the tensor completion problem efficiently. However, these models fail to exploit the local patterns of tensors, which compromises the performance of tensor completion. In this paper, we propose a novel hierarchical prior regularized matrix factorization model for tensor completion. This model hierarchically incorporates the low-rank prior, total variation prior, and sparse coding prior into a matrix factorization, simultaneously characterizing both the global low-rank property and the local smoothness of tensors. For solving the proposed model, we use the alternating direction method of multipliers to establish our algorithm. Besides, the complexity and convergence are investigated to further validate the algorithm effectiveness. The proposed scheme is then evaluated through various data sets. Experiment results verify that, the proposed method outperforms several state-of-the-art approaches.