Abstract
Utilizing separation to decompose a local filter mask is a well-known technique to accelerate its convolution with discrete two-dimensional signals such as images. However, many modern days’ applications involve higher-dimensional, discrete data that needs to be processed but whose inherent spatial complexity would render immediate/naive convolutions computationally infeasible. In this paper, we show how separability of general higher-order tensors can be leveraged to reduce the computational effort for discrete convolutions from super-polynomial to polynomial (in both the filter mask’s tensor order and spatial expansion). Thus, where applicable, our method compares favorably to current tensor convolution methods and, it renders linear filtering applicable to signal domains whose spatial complexity would otherwise have been prohibitively high. In addition to our theoretical guarantees, we experimentally illustrate our approach to be highly beneficial not only in theory but also in practice.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Explore related subjects
Discover the latest articles and news from researchers in related subjects, suggested using machine learning.References
Beylkin, G., Garcke, J., & Mohlenkamp, M. J. (2008). Multivariate regression and machine learning with sums of separable function. In SIAM Journal on Scientific Computing.
Bro, R. (1997). PARAFAC. Tutorial and applications. Chemometrics and Intelligent Laboratory Systems, 149–171.
Carroll, J. D., & Chang, J. J. (1970). Analysis of individual differences in multidimensional scaling via an n-way generalization of “eckart-young” decomposition. In Psychometrika, 283–319.
Comon P. (2002) Mathematics in Signal Processing V, Chapter Tensor Decompositions. Clarendon Press, Oxford, UK, pp 1–24
Deng, T., Kawamata, M., & Higuchi, T. (1991) Design of two-dimensional recursive digital filters based on the optimal decomposition of magnitude specifications. In Circuits and Systems, IEEE International Symposium on, 614–617.
Deriche, R. (1987). Using canny’s criteriato derivea recursively implemented optimal edge detector. In International Journal of Computer Vision.
Deriche, R. (1992). Recursively implementing the gaussian and its derivatives. In: Proceedings of the 2nd International Conference on Image Processing, 263–267, Singapore.
Derpanis, K., & Gryn, J. (2004) Three-dimensional nth derivative of gaussian separable steerable filters. In Technical Report CS-2004-05, York University.
Dunlavy, D., Kolda, T. G., & Kegelmeyer, P. (2008). Tensor decomposition for analyzing multi-link graphs. In SIAM Parallel Processing for Scientific Computing.
Farnebäck, G., & Westin, C. F. (2006). Improving deriche-style recursive gaussian filters. Journal of Mathematical Imaging and Vision, 293–299.
Gibson, C. J. (1993). Rapid filtering of large medical images using one-dimensional convolution kernels. Computer Biology and Medicine, 265–272.
Hale, D. (2006). Recursive gaussian filters. Center for Wave Phenomena Report 546.
Harshman, R. A. (1970). Foundations of the PARAFAC procedure: Models and conditions for an “explanatory” multi-modal factor analysis. In UCLA Working Papers in Phonetics, 1–84.
Hitchcock, F. L. (1927). The Expression of a tensor or a polyadic as a sum of products. In Journal of Mathematics and Physics, 164–189.
Ibanez, L., Schroeder, W., Ng, L., & Cates, J. (2005). The ITK Software Guide. Albany, NY: Kitware Inc. http://www.itk.org/ItkSoftwareGuide.pdf.
Khoromskij, B. N. (2010). Fast and accurate tensor approximation of a multivariate convolution with linear scaling in dimension. Journal of Computation and Applied Mathematics.
Kolda, T. G., & Bader, B. W. (2009). Tensor decompositions and applications. In SIAM Review, 455–500.
Kruger, R. P., Hall, E. L., Dwyer, S. J., & Lodwick, G. S. (1971). Digital techniques for image enhancement of radiographs. In International Journal of Bio-Medical Computing, 215–238.
Lathauwer, L. d., Moor, B. d., & Vandewalle, J. (2000). On the best rank–1 and rank − (R 1, R 2, . . . , R n ) approximation and applications of higher-order tensors. In SIAM Journal on MATRIX Analysis and Applications, 1324–1342.
Lathauwer, L. d., & Vandewalle, J. (2004). Dimensionality reduction in higher-order signal processing and rank − (R 1, R 2, . . . , R N ) reduction in multilinear algebra. In Linear Algebra and its Applications, 31–55.
Luong, Bruno (2009). FFT-based convolution (convnfft). http://www.mathworks.com/matlabcentral/fileexchange/24504-fft-based-convolution, Release: 16 Sep 2009.
Monga O. (1988). Image segmentation by regions hierarchical growth. In 09—Electronics and Electrical Engineering, Computer Science.
Muti, D., & Bourenname, S. (2005). Multidimensional filtering based on a tensor approach. In Signal Processing.
Pawley, J. B. (2008). Handbook of biological confocal microscopy. Journal of Biomedical Optics.
Pfeiffer, S., Mai, M., Globke, W., & Calliess, J. (2009). On generalized separation and the speedup of local operators on multi-dimensional signals. In Multidimensional (nD) Systems, 2009, 1–8.
Schuster, D. (1990). Separierbarkeit zweidimensionaler filter. In 12. DAGM-Symposium, Informatik- Fachberichte, 288–291.
Svensson, B. (2008). A Multidimensional Filtering Framework with Applications to Local Structure Analysis and Image Enhancement. PhD thesis, Sweden: Linköping University.
Treitel, S., & Shanks, J. L. (1971). The design of mutistage separable planar filters. In Geoscience Electronics, IEEE Transactions on, 10–27.
Twogood, R., & Mitra, S. (1977). Computer-aided design of separable two-dimensional digital filters. In Acoustics, Speech and Signal Processing, IEEE Transactions on, 165–169.
Vos, M. d., Vergult, A., Lathauwer, L. d., Clercq, W. d., Huffel, S. v., Dupont, Paomini, P. A., & Paesschen, W. v. (2007). Canonical decomposition of ictal scalp EEG reliably detects the seizure onset zone. In NeuroImage, 844–854.
Vliet, L. v., Young, I., & Verbeek, P. (1998) Recursive gaussian derivative filters. In: Proceedings of the International Conference on Pattern Recognition, (pp. 509–514). Brisbane.
Yener, B., Evrim, A., Aguis, P., Bennett, K., Vandenberg, S. L., & Plopper, G. E. (2008). Multiway modeling and analysis in stem cell systems biology. BiodMed Central Systems Biology, 17.
Zhang, P. J., Passmore, Y., & Bayford, R. H. (2005). Visualization and Post-processing of 5D brain images. In Engineering in Medicine and Biology 27th Annual Conference, 1083–1086.
Zhang, T., & Golub, G. H. (2001). Rank-one approximation to high order tensors. SIAM, Journal on Matrix Analysis and Applications, 534–550.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Calliess, JP., Mai, M. & Pfeiffer, S. On the computational benefit of tensor separation for high-dimensional discrete convolutions. Multidim Syst Sign Process 23, 255–279 (2012). https://doi.org/10.1007/s11045-010-0131-2
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11045-010-0131-2