Problem of minimizing a sum of differences of weighted convolutions
In this paper the problem of minimizing a sum of differences of weighted convolutions is formulated as the problem of optimal summing of elements of two sequences where indices play the role of variables. It is shown that the considered problem can be interpreted as the problem that minimizes the sum of squared distances between the elements of an input sequence and one of the admissible approximating sequences. The considered problem is a generalization of the problems investigated in [\textit{A. V. Kel'manov} et al., Sib. Zh. Ind. Mat. 5, No. 1, 94--108 (2002; Zbl 1075.93539)]. The size of the set of all feasible solutions is estimated. Using a dynamic programming approach and the Bellman's principle of optimality an exact polynomial algorithm solving the considered problem is obtained. The proposed algorithm is tested on two examples of numerical simulation of processing medical signals, namely electrocardiogram (ECG)-like, and photoplethysmogram (PPG)-like pulses distorted by additional noise.
- One problem of the sum of weighted convolution differences minimization, induced by the quasiperiodic sequence recognition problem
- scientific article; zbMATH DE number 14805
- Weighted averaging using adaptive estimation of the weights
- Fast algorithms for the maximum convolution problem
- Publication:4725230
This page was built for publication: Problem of minimizing a sum of differences of weighted convolutions
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q828127)