Three-Mode Biomedical Signal Denoising in the Local Phase Space based on a Tensor Approach

  • K. Pukenas


In this research, an improved local projection noise reduction approach with three-mode model of neighborhood is proposed. Firstly, one dimensional time series are embedded into a high dimensional phase space. Secondly, the neighborhood tensor of each reference no overlapping window with several consecutive vectors of reconstructed phase-space is computed rather than neighborhood matrix of each separate vector. Lastly, with the suggested model a higher order singular value decomposition (HOSVD) is performed on the neighborhood tensors to split the three mode data into two orthogonal subspaces: the signal and noise subspaces. Throughout the experiment, the effectiveness of the proposed method is validated with a noisy simulated data — the x component of Rossler system and real biomedical signal contaminated with additive white Gaussian noise. Ill. 3, bibl. 14 (in English; abstracts in English and Lithuanian).