Precise deconvolution using the Fermat number transform

From MaRDI portal





The subject of the paper is the description of an error-free algorithm to solve convolution systems of linear equations. The results described in the paper show that the Fermat number transform and modular arithmetic can be successfully used, not only in solving the problems of fast precise linear convolution filters but also in solving the problems of deconvolution. The described algorithm differs from all deconvolution algorithms known so far. The main point of the usage modular arithmetic in this field consists in an elimination of the rounding-off errros. The commonly used Fourier transform unavoidably gives a solution with oscillations as these inverse problems are extremely sensitive to this type of error. A fast algorithm to calculate the determinant of the convolution matrix is also given as its exact integer value is needed for the deconvolution algorithm. The described deconvolution method is used to the resolution enhancement in processing of spectroscopic data measured in nuclear physics experiments.











This page was built for publication: Precise deconvolution using the Fermat number transform

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3750038)