Image-Based Process Monitoring Using Low-Rank Tensor Decomposition

Yan, Hao; Paynabar, Kamran; Shi, Jianjun

doi:10.1109/tase.2014.2327029

Cited by 112 publications

(47 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The use of image data for part quality inspection is becoming more widespread in industry (Qiu, 2005;Yan et al, 2015;Megahed et al, 2011). Random porous and cellular materials represent a category of complicated structures where novel SPC methods are needed to cope with the challenging nature of quality inspection data (Kim et al, 2014;Zhuravleva et al, 2013;Campoli et al, 2013;Banhart, 2001).…”

Section: A Motivating Real Case Studymentioning

confidence: 99%

Profile Monitoring of Probability Density Functions via Simplicial Functional PCA With Application to Image Data

et al. 2018

View full text Add to dashboard Cite

The advance of sensor and information technologies is leading to data-rich industrial environments, where big amounts of data are potentially available. In this scenario, image data play a relevant role, as they can easily describe many phenomena of interest. This study focuses on images where several and similar features of interest are randomly distributed and characterized by no spatially correlated structure. Examples are pores in parts obtained via casting or additive manufacturing, voids in metal foams and light-weight components, grains in metallographic analysis, etc. The proposed approach consists of summarizing the random occurrences of the observed features via its (empirical) probability density function (PDF). In particular, a novel approach for PDF monitoring is proposed. It is based on simplicial functional principal component analysis (SFPCA), which is performed by applying an isometric isomorphism between the space of density functions, i.e., the Bayes space B 2 , and the space of square integrable functions L 2 . A simulation study shows the enhanced monitoring performances provided by the SFPCA-based profile monitoring against other competitors proposed in the literature. Eventually, a real case study dealing with the quality control of foamed materials production is discussed, to highlight a practical use of the proposed methodology.

show abstract

Section: A Motivating Real Case Studymentioning

confidence: 99%

Profile Monitoring of Probability Density Functions via Simplicial Functional PCA With Application to Image Data

et al. 2018

View full text Add to dashboard Cite

show abstract

“…While tensor voting in images has been extensively studied [7], [45], there is little work on tensor voting of 3-D surfaces. This is despite the fact that the 3-D surfaces become ever more widespread and have myriad applications.…”

Section: G Tensor Votingmentioning

confidence: 99%

Tensor Voting Guided Mesh Denoising

Wei

Liang

Pang

et al. 2017

IEEE Trans. Automat. Sci. Eng.

View full text Add to dashboard Cite

Mesh denoising is imperative for improving imperfect surfaces acquired by scanning devices. The main challenge is to faithfully retain geometric features and avoid introducing additional artifacts when removing noise. Unlike the existing mesh denoising techniques that focus only on either the first-order features or high-order differential properties, our approach exploits the synergy when facet normals and quadric surfaces are integrated to recover a piecewise smooth surface. In specific, we vote on surface normal tensors from robust statistics to guide the creation of consistent subneighborhoods subsequently used by moving least squares (MLS). This voting naturally leads to a conceptually simple way that gives a unified mesh-denoising framework for not only handling noise but also enabling the recovering of surfaces with both sharp and small-scale features. The effectiveness of our framework stems from: 1) the multiscale tensor voting that avoids the influence from noise; 2) the effective energy minimization strategy to searching the consistent subneighborhoods; and 3) the piecewise MLS that fully prevents the side effects from different subneighborhoods during surface fitting. Our framework is direct, practical, and easy to understand. Comparisons with the state-of-the-art methods demonstrate its outstanding performance on feature preservation and artifact suppression.Note to Practitioners-Three-dimensional sensing and scanning devices are widely used to capture digital surfaces of real objects and scenes in many scenarios. However, due to occlusion, motion, multiple reflections, and so on, the captured data often suffer from severe contamination with noise, significantly hindering its practical applications. Therefore, it is indispensable to remove noise prior to further processing, which is commonly referred to as mesh denoising. Mesh denoising is a long-standing problem, and remains open in academic as well as industrial applications due to its challenging nature. The state-of-the-art methods either fail to retain most of the original features presented well in the object, or cannot avoid additional artifacts, such as vertex drifts. In contrast, we design a denoising framework aiming at improving the quality of the raw surface by producing a mesh with better perceptual features. The technique developed here can produce high-quality surface data of real objects and scenes, which would facilitate the modeling, reconstruction, and recognition applications in computer-aided design, reverse engineering, 3-D printing, and computer-aided manufacturing.Index Terms-Consistent subneighborhood, feature preserving, mesh denoising, multiscale tensor voting, piecewise moving least squares (pMLS).

show abstract

“…Morphology of nanoparticles was characterized by a multistage procedure and then semiautomatically classify them into homogeneous groups. Yan et al proposed to integrate low‐rank tensor decomposition with multivariate control charts for image‐based process monitoring . In addition, Zhang et al measured the variations of wafer thickness from image profiles using an adaptive Gaussian process model.…”

Section: Research Backgroundmentioning

confidence: 99%

“…Yan et al proposed to integrate low-rank tensor decomposition with multivariate control charts for imagebased process monitoring. 23 In addition, Zhang et al 3 measured the variations of wafer thickness from image profiles using an adaptive Gaussian process model.…”

Section: Image Profilesmentioning

confidence: 99%

Dynamic network monitoring and control of in situ image profiles from ultraprecision machining and biomanufacturing processes

Chen

Yang

2017

Quality & Reliability Eng

View full text Add to dashboard Cite

In modern industries, advanced imaging technology has been more and more invested to cope with the ever‐increasing complexity of systems, to improve the visibility of information and enhance operational quality and integrity. As a result, large amounts of imaging data are readily available. This presents great challenges on the state‐of‐the‐art practices in process monitoring and quality control. Conventional statistical process control (SPC) focuses on key characteristics of the product or process and is rather limited to handle complex structures of high‐dimensional imaging data. New SPC methods and tools are urgently needed to extract useful information from in situ image profiles for process monitoring and quality control. In this study, we developed a novel dynamic network scheme to represent, model, and control time‐varying image profiles. Potts model Hamiltonian approach is introduced to characterize community patterns and organizational behaviors in the dynamic network. Further, new statistics are extracted from network communities to characterize and quantify dynamic structures of image profiles. Finally, we design and develop a new control chart, namely, network‐generalized likelihood ratio chart, to detect the change point of the underlying dynamics of complex processes. The proposed methodology is implemented and evaluated for real‐world applications in ultraprecision machining and biomanufacturing processes. Experimental results show that the proposed approach effectively characterize and monitor the variations in complex structures of time‐varying image data. The new dynamic network SPC method is shown to have strong potentials for general applications in a diverse set of domains with in situ imaging data.

show abstract

Image-Based Process Monitoring Using Low-Rank Tensor Decomposition

Cited by 112 publications

References 33 publications

Profile Monitoring of Probability Density Functions via Simplicial Functional PCA With Application to Image Data

Profile Monitoring of Probability Density Functions via Simplicial Functional PCA With Application to Image Data

Tensor Voting Guided Mesh Denoising

Dynamic network monitoring and control of in situ image profiles from ultraprecision machining and biomanufacturing processes

Contact Info

Product

Resources

About