Autors: Kanev, I. T., Pavlova, P. E., Kromichev D., Genoff, J. G.
Title: Generating 2D Gaussian Kernels for the Purpose of Studying FPGA-Based Linear Spatial Filters
Keywords: linear spalial filters, LSF, weighted average, normalized Gaussian 2D Kernel, VHDL, FPGA

Abstract: This paper presents one approach to generating normalized 2D Gaussian kernels. Defined are the conditions for determining the coefficients of the kernels. Selected is an approach through which the kernel coefficients are represented as positive integers. This approach aims at speeding up the time to calculate the linear spatial fillers (LSF) and reducing the number of logic elements. An algorithm for generating 2D Gaussian kernels is proposed. The studies are conducted using the software tool "Gaussian Kernel Generator (GKG)". As a result of the studies, the Gaussian kernel coefficients, which are represented as real and integer numbers, are determined on the basis of the preset criteria. The obtained results can be employed in designing and studying LSF through using FPGA platforms.



    Journal of Technical University Sofia, Plovdiv branch ”Fundamental Sciences and Applications“, pp. 111-118, 2016, Bulgaria, ISBN ISSN 1310-8271

    Вид: статия в списание