Online Tensor Robust Principal Component Analysis

Salut, Mohammad Mohammadpour; Anderson, David V.

doi:10.1109/access.2022.3186364

Cited by 10 publications

(4 citation statements)

References 63 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Another notable application of tensor tracking in computer vision is video background and foreground separation which is quite related to visual tracking, but with a different aim of modeling the scene background and detecting the information of changes in the scene. Similar to visual tracking, many tensorbased separators were proposed, such as [65], [105], [112], [124], [125]. Particularly in [65], Thanh et al proposed a robust adaptive CP method called RACP which is capable of modeling video background and detecting moving objects.…”

Section: Computer Visionmentioning

confidence: 99%

See 1 more Smart Citation

A Contemporary and Comprehensive Survey on Streaming Tensor Decomposition

Thanh¹,

abed-meraim²,

Linh-Trung³

et al. 2022

Preprint

View full text Add to dashboard Cite

<p>Tensor decomposition has been demonstrated to be successful in a wide range of applications, from neuroscience and wireless communications to social networks. In an online setting, factorizing tensors derived from multidimensional data streams is however non-trivial due to several inherent problems of real-time stream processing. In recent years, many research efforts have been dedicated to developing online techniques for decomposing such tensors, resulting in significant advances in streaming tensor decomposition or tensor tracking. This topic is emerging and enriches the literature on tensor decomposition, particularly from the data stream analystics perspective. Thus, it is imperative to carry out an overview of tensor tracking to help researchers and practitioners understand its development and achievements, summarise the current trends and advances, and identify challenging problems. In this article, we provide a contemporary and comprehensive survey on different types of tensor tracking techniques. We particularly categorize the state-of-the-art methods into three main groups: streaming CP decompositions, streaming Tucker decompositions, and streaming decompositions under other tensor formats (i.e., tensor-train, t-SVD, and BTD). In each group, we further divide the existing algorithms into sub-categories based on their main optimization framework and model architectures. Finally, we present several research challenges, open problems, and potential directions of tensor tracking in the future.</p>

show abstract

Section: Computer Visionmentioning

confidence: 99%

“…A robust streaming tensor-train algorithm was developed in [112] which also has the potential to detect foreground in video. Salut et al in [125] proposed an online tensor robust principal component analysis and validated its effectiveness with the problem of background and foreground separation.…”

Section: Computer Visionmentioning

confidence: 99%

A Contemporary and Comprehensive Survey on Streaming Tensor Decomposition

Thanh¹,

abed-meraim²,

Linh-Trung³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…The goal of data compression algorithms is to reduce the volume of information features as much as possible to ensure that the information is more complete, saving costs; due to the amount of data and the diversity of structures, traditional data compression algorithms are faced with computationally inefficient as well as the waste of storage resources and other problems [15][16]. Literature [17] explored the compression method of GPU register file data in order to improve the parallel efficiency of GPU registers through data compression to reduce the width of register file read and write operations.…”

Section: Introductionmentioning

confidence: 99%

Compression of electrical code violation recognition data using the improved swinging door trending algorithm

Yang,

Zhao,

Han

et al. 2024

Applied Mathematics and Nonlinear Sciences

View full text Add to dashboard Cite

Aiming at the challenge of storing massive power grid data, this paper proposes an improved swing gate trend algorithm to effectively compress 5G data. The algorithm first performs least squares smoothing on the original data to reduce noise interference on the SDT algorithm, which enables the data compression process to more accurately determine the data trend. Further, the shortcomings of the original SDT algorithm are improved, including adaptive frequency conversion data processing, dynamic threshold adjustment, and anomaly recording strategy, to enhance the practicality and efficiency of the algorithm. Through simulation analysis and example data validation, the study shows that the data compression ratio can be stabilized at about 23.98 when the data compression time reaches 1.6 minutes, and the actual error is very close to the desired error. The time overhead of the improved SDT algorithm is only 0.225 seconds, indicating that the algorithm is efficient and reliable. Combined with different data compression storage strategies, the algorithm can further reduce the data compression time. This study provides an adequate data compression method for electric code violation identification, which offers a practical solution for processing and storing large-scale grid data.

show abstract

“…Despite having advantages, most of the existing streaming Tucker decomposition algorithms mentioned above are sensitive to sparse outliers. In the adaptive signal processing literature, there are some other streaming tensor methods robust to data corruptions such as in [31]- [34]. However, they are specifically designed under other tensor formats (i.e., CP/PARAFAC, tensor-train, and t-SVD).…”

Section: Introductionmentioning

confidence: 99%

Robust Online Tucker Dictionary Learning from Multidimensional Data Streams

Thanh

Duy

Abed-Meraim

et al. 2022

2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)

View full text Add to dashboard Cite

Big data streaming analytics has recently attracted much attention in the signal and information processing communities due to the fact that massive streaming datasets have been collected over the years. Among them, many modern data streams are represented as multidimensional arrays (aka tensors), and thus, streaming tensor decomposition or tensor tracking has become a promising tool to analyze such streaming data. In this paper, we propose a novel online algorithm called ROTDL for the problem of robust tensor tracking under the Tucker format. ROTDL is not only capable of tracking the underlying Tucker dictionary of multidimensional data streams over time, but also robust to sparse outliers. The proposed algorithm is specifically designed by using the alternating direction method of multipliers, block-coordinate descent, and recursive least-squares filtering techniques. Several experiments demonstrate the effectiveness of ROTDL for robust tensor tracking.

show abstract

Online Tensor Robust Principal Component Analysis

Cited by 10 publications

References 63 publications

A Contemporary and Comprehensive Survey on Streaming Tensor Decomposition

A Contemporary and Comprehensive Survey on Streaming Tensor Decomposition

Compression of electrical code violation recognition data using the improved swinging door trending algorithm

Robust Online Tucker Dictionary Learning from Multidimensional Data Streams

Contact Info

Product

Resources

About