1. Tensor ring decomposition for visual data denoising via tensor random projection
Authors : Longhao Yuan, Jianting Cao, Qibin Zhao
Pages : 102-107
DOI : http://dx.doi.org/10.21172/1.132.19
Keywords : Large-scale datatensor decompositiontensor ring decompositionrandomized algorithmdenoising Abstract :Abstract- Large-scale data have posed a big computational challenge to traditional data processing methods. The recently proposed tensor ring decomposition (TRD) has shown to be a promising tool to process large-scale data. However, the existing TRD algorithms are of high computational cost which makes it less efficient to apply to large-scale datasets. In this paper, by employing the random projection method to TRD, we propose a non-iterative TRD algorithm for fast large-scale data decomposition which can be used for image denoising tasks. In the experiments of large visual data denoising, our method shows satisfying performance and huge speed-up without loss of accuracy, in comparison with traditional algorithms.
Citing this Journal Article :Longhao Yuan, Jianting Cao, Qibin Zhao, "Tensor ring decomposition for visual data denoising via tensor random projection", Volume 13 Issue 2 - April 2019, 102-107
Click here to Submit Copyright Takedown Notice for this article.