Abstract
Tensor representation of video sequences or 3D images becomes very popular recently. Despite certain advantages, the main obstacle for the wide application of this approach is the high computational complexity. In the paper is presented new method for third-order tensor representation in the spectrum space of the 3D-Walsh–Hadamard Transform (WHT). To lessen the computational complexity, here is used pyramidal decomposition based on the 3D WHT with Reduces Inverse Spectrum Pyramid (RISP). In result is achieved high concentration of the tensor energy in a minimum number of spectrum coefficients, most of which—in the first (lowest) decomposition level. After the processing, the tensor is transformed into a multi-level spectrum tensor of same size. The proposed representation has low computational complexity because its execution needs operations like “addition” only. The specific properties of the 3D-RISP permit it to be used in various application areas which require parallel processing and analysis of 3D data represented as third-order tensors: sequences of correlated images (video, multi-spectral, multi-view, various kinds of medical images), multichannel signals, huge massifs of data, etc.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
References
Hackbusch, W., Kühn, S.: A new scheme for the tensor representation. J. Fourier Anal. Appl. 15(5), 706–722 (2009)
De Lathauwer, L., De Moor, B., Vandewalle, J.: A multilinear singular value decomposition. SIAM J. Matrix Anal. Appl. 21(4), 1253–1278 (2000)
Lu, H., Plataniotis, K., Venetsanopulos, A.: Multilinear principal component analysis of tensor objects. IEEE Trans. Neural Netw. 19(1), 18–39 (2008)
Domanov, I., De Lathauwer, L.: Canonical polyadic decomposition of third-order tensors: relaxed uniqueness conditions and algebraic algorithm. SIAM J. Matrix Anal. Appl. Soc. Ind. Appl. Math. 35(2), 636–660 (2014)
Kolda, T., Bader, B.: Tensor decompositions and applications. SIAM Rev. 51(3), 455–500 (2009)
He, Z., Gao, S., Xiao, L., Liu, D., He, H., Barber, D.: Wider and Deeper, Cheaper and Faster: Tensorized LSTMs for sequence learning. In: 31st Conference on Neural Information Processing Systems (NIPS 2017), CA, USA, pp. 1–11 (2017)
Zhang, A., Xia, D.: Tensor SVD: statistical and computational limits. IEEE Trans. Inf. Theory 64(11), 7311–7338 (2018)
Grasedyck, L.: Hierarchical singular value decomposition of tensors. SIAM J. Matrix Anal. Appl. 31(4), 2029–2054 (2010)
Bergqvist, G., Larsson, E.: The higher-order singular value decomposition: theory and an application. IEEE Signal Process. Mag. 27(3), 151–154 (2010)
Cichocki, A., Mandic, D., Phan, A., Caiafa, C., Zhou, G., Zhao, Q., De Lathauwer, L.: Tensor decompositions for signal processing applications: from two-way to multi-way component analysis. IEEE Signal Process. Mag. 32(2), 145–163 (2015)
Sakai, T., Sedukhin, S.: 3D discrete transforms with cubical data decomposition on the IBM Blue Gene/Q. Technical Report 2013–001, Graduate School of Computer Science and Engineering, The University of Aizu, 31 p. (2013)
Agaian, S., Sarukhanyan, H., Egiazarian, K., Astola, J.: Hadamard Transforms. SPIE Press Book, Washington (2011)
Kountchev, R., Kountcheva, R.: Image representation with reduced spectrum pyramid. In: Tsihrintzis, G., Virvou, M., Howlett, R., Jain, L. (eds.) Chapter in: New Directions in Intelligent Interactive Multimedia, pp. 275–284. Springer (2008)
Kountchev, R., Rubin, S., Milanova, M., Kountcheva, R.: Comparison of image decompositions through Inverse difference and Laplacian pyramids. J. Multimed. Data Eng. Manag. 6(1), 19–38 (2015)
Kountchev, R., Kountcheva, R.: Hierarchical decomposition of 2D/3D images, based on SVD2 × 2. Int. J. Multimed. Image Process. (IJMIP) 5(3/4), 286–296 (2015)
Kountchev, R., Kountcheva, R.: Image sequence decomposition based on the Truncated Hierarchical SVD. Int. J. Multimed. Image Process. (IJMIP) 7(1), 352–361 (2017)
Acknowledgements
This work was supported by the National Science Fund of Bulgaria: Project No. KP-06-H27/16 “Development of efficient methods and algorithms for tensor-based processing and analysis of multidimensional images with application in interdisciplinary areas”.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Kountchev, R., Kountcheva, R. (2020). Low Computational Complexity Third-Order Tensor Representation Through Inverse Spectrum Pyramid. In: Kountchev, R., Patnaik, S., Shi, J., Favorskaya, M. (eds) Advances in 3D Image and Graphics Representation, Analysis, Computing and Information Technology. Smart Innovation, Systems and Technologies, vol 179. Springer, Singapore. https://doi.org/10.1007/978-981-15-3863-6_8
Download citation
DOI: https://doi.org/10.1007/978-981-15-3863-6_8
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-15-3862-9
Online ISBN: 978-981-15-3863-6
eBook Packages: Intelligent Technologies and RoboticsIntelligent Technologies and Robotics (R0)