Additive Tensor Decomposition Considering Structural Data Information

Tensor data with rich structural information becomes increasingly important in process modeling, monitoring, and diagnosis. Here structural information is referred to structural properties such as sparsity, smoothness, low-rank, and piecewise constancy. To reveal useful information from tensor data, we propose to decompose the tensor into the summation of multiple components based on different structural information of them. In this paper, we provide a new definition of structural information in tensor data. Based on it, we propose an additive tensor decomposition (ATD) framework to extract useful information from tensor data. This framework specifies a high dimensional optimization problem to obtain the components with distinct structural information. An alternating direction method of multipliers (ADMM) algorithm is proposed to solve it, which is highly parallelable and thus suitable for the proposed optimization problem. Two simulation examples and a real case study in medical image analysis illustrate the versatility and effectiveness of the ATD framework.

[1]  J. Ramsay.  Monotone Regression Splines in Action , 1988 .

[2]  F. O’Sullivan.  Discretized Laplacian Smoothing by Fourier Methods , 1991 .

[3]  D. Ruppert.  The Elements of Statistical Learning: Data Mining, Inference, and Prediction , 2004 .

[4]  R. Tibshirani,et al.  Sparsity and smoothness via the fused lasso , 2005 .

[5]  Tamara G. Kolda,et al.  Tensor Decompositions and Applications , 2009, SIAM Rev..

[6]  Robert Tibshirani,et al.  The Elements of Statistical Learning: Data Mining, Inference, and Prediction, 2nd Edition , 2001, Springer Series in Statistics.

[7]  Yi Ma,et al.  Robust principal component analysis? , 2009, JACM.

[8]  Yi Ma,et al.  TILT: Transform Invariant Low-Rank Textures , 2010, ACCV 2010.

[9]  T. Jayakumar,et al.  Infrared thermography for condition monitoring – A review , 2013 .

[10]  Thierry Bouwmans,et al.  Robust PCA via Principal Component Pursuit: A review for a comparative evaluation in video surveillance , 2014, Comput. Vis. Image Underst..

[11]  A. Einstein,et al.  Quantity and location of aortic valve complex calcification predicts severity and location of paravalvular regurgitation and frequency of post-dilation after balloon-expandable transcatheter aortic valve replacement. , 2014, JACC. Cardiovascular interventions.

[12]  Stephen P. Boyd,et al.  Proximal Algorithms , 2013, Found. Trends Optim..

[13]  Hao Yan,et al.  Image-Based Process Monitoring Using Low-Rank Tensor Decomposition , 2015, IEEE Transactions on Automation Science and Engineering.

[14]  Z. Qian,et al.  Quantitative Prediction of Paravalvular Leak in Transcatheter Aortic Valve Replacement Based on Tissue-Mimicking 3D Printing. , 2017, JACC Cardiovascular Imaging.

[15]  Hao Yan,et al.  Anomaly Detection in Images With Smooth Background via Smooth-Sparse Decomposition , 2017, Technometrics.

[16]  Piotr Duda,et al.  Identification of overheating of an industrial fluidized catalytic cracking regenerator , 2018 .

[17]  Sumathi Poobal,et al.  Crack detection using image processing: A critical review and analysis , 2017, Alexandria Engineering Journal.

[18]  Hao Yan,et al.  Real-Time Monitoring of High-Dimensional Functional Data Streams via Spatio-Temporal Smooth Sparse Decomposition , 2018, Technometrics.

[19]  Nhat-Duc Hoang,et al.  Detection of Surface Crack in Building Structures Using Image Processing Technique with an Improved Otsu Method for Image Thresholding , 2018 .

[20]  Kan Wang,et al.  An efficient statistical approach to design 3D-printed metamaterials for mimicking mechanical properties of soft biological tissues , 2018, Additive Manufacturing.