Superresolution from Principal Component Models by RKHS Sampling

  • J. P. Lewis
  • Ken Anjyo
  • Taehyun Rhee
Conference paper
Part of the Mathematics for Industry book series (MFI, volume 18)


Principal component analysis (PCA) involves a signal that is sampled at some arbitrary but fixed and finite set of locations. Radial Basis Function (RBF) regression interpolates a-priori known data to arbitrary locations as a weighted sum of a (radial) kernel function centered at the data points. In recent work we showed that if the RBF kernel is equated to the covariance, RBF and Gaussian Process (GP) models perform a similar computation, differing in what information is assumed known in advance, and what is known at runtime. Building on the RBF-GP equivalence, we show that if the data covariance is known (or can be estimated), an RBF-inspired regression can provide data-driven “superresolution” interpolation of given data. This procedure can alternately be interpreted as a superresolution extension of eigenvector (principal component) data models, as signal sampling (function evaluation) in a discrete reproducing kernel Hilbert space (RKHS) generated by the data covariance, or as an elementary Gaussian process model in which the observations have a low-rank representation.


Radial basis function regression Principal component analysis  Gaussian process 



This research is partially supported by the Japan Science and Technology Agency, CREST project. We would like to thank the reviewers, Hiroyuki Ochiai, and Ayumi Kimura. Data used in this project was obtained from, created with funding from NSF EIA-0196217 [5].


  1. 1.
    K. Anjyo, J. Lewis, RBF interpolation and Gaussian process regression through an RKHS formulation. JMI 3, 63–71 (2011)Google Scholar
  2. 2.
    K. Anjyo, J. Lewis, Scattered data interpolation for computer graphics. SIGGRAPH Course. (2014)
  3. 3.
    J.C. Carr, R.K. Beatson, J.B. Cherrie, T.J. Mitchell, W.R. Fright, B.C. McCallum, T.R. Evans, Reconstruction and representation of 3D objects with radial basis functions, in SIGGRAPH’01: proceedings of the 28th annual conference on Computer graphics and interactive techniques, (ACM, New York, NY, USA, 2001) pp. 67–76. doi:
  4. 4.
    G. Casciola, L. Montefusco, S. Morigi, Edge-driven image interpolation using adaptive anisotropic radial basis functions. J. Math. Imaging Vis. 36(2), 125–139 (2010)MathSciNetCrossRefGoogle Scholar
  5. 5.
    CMU Graphics Lab Motion Capture Database. (2015)
  6. 6.
    T.F. Cootes, G.J. Edwards, C.J. Taylor, Active appearance models, Lecture Notes in Computer Science, vol. 1407 (Springer, Berlin, 1998)Google Scholar
  7. 7.
    H.Q. Dinh, G. Turk, G. Slabaugh, Reconstructing surfaces using anisotropic basis functions, in Proceedings of Eighth IEEE International Conference on Computer Vision, ICCV 2001, IEEE, vol. 2, pp. 606–613 (2001)Google Scholar
  8. 8.
    J. Gentle, Random Number Generation and Monte Carlo Methods, Statistics and Computing (Springer, New York, 2003)Google Scholar
  9. 9.
    P.J. Green, B.W. Silverman, Nonparametric Regression and Generalized Linear Models: A Roughness Penalty Approach (Chapman and Hall, London, 1994)zbMATHCrossRefGoogle Scholar
  10. 10.
    M. Pourahmadi, High-Dimensional Covariance Estimation: With High-Dimensional Data (Wiley, Wiley Series in Probability and Statistics, 2013)CrossRefGoogle Scholar
  11. 11.
    K. Rao, P. Yip, Discrete Cosine Transform: Algorithms, Advantages, Applications (Academic Press, San Diego, 1990)Google Scholar
  12. 12.
    C. Rasmussen, C. Williams, Gaussian Processes for Machine Learning (MIT Press, Adaptive Computation and Machine Learning, 2006)zbMATHGoogle Scholar
  13. 13.
    M. Salzmann, R. Urtasun, Implicitly constrained Gaussian process regression for monocular non-rigid pose estimation, in Advances in Neural Information Processing Systems, vol. 23, ed. by J. Lafferty, C. Williams, J. Shawe-Taylor, R. Zemel, A. Culotta (Curran Associates, Inc., 2010) pp. 2065–2073Google Scholar
  14. 14.
    Y. Seol, J. Seo, P.H. Kim, J.P. Lewis, J. Noh, Weighted pose space editing for facial animation. Vis. Comput. 28(3), 319–327 (2012)CrossRefGoogle Scholar
  15. 15.
    T. Valentine, Face-space models of face recognition, in Computational, Geometric, and Process Perspectives on Facial Cognition: Contexts and Challenges, Scientific Psychology Series, ed. by M. Wenger, J. Townsend (Taylor and Francis, 2012)Google Scholar

Copyright information

© Springer Japan 2015

Authors and Affiliations

  1. 1.Victoria University and Weta Digital/JST CRESTWellingtonNew Zealand
  2. 2.OLM Digital/JST CRESTSetagaya-ku, TokyoJapan
  3. 3.Victoria UniversityWellingtonNew Zealand

Personalised recommendations