학술논문
Separable convolution gaussian smoothing filters on a xilinx FPGA platform
Document Type
Conference
Author
Source
Fifth International Conference on the Innovative Computing Technology (INTECH 2015) Innovative Computing Technology (INTECH), 2015 Fifth International Conference on. :112-117 May, 2015
Subject
Language
Abstract
Convolution Gaussian filtering is a technique that can be used for a wide array of image processing tasks, some of which may include smoothing, edge detection, line parameter estimation and texture analysis. The greatest advantage of the Gaussian filters is that they are separable. Our main goal in this paper is to develop an efficient architecture for Separable Gaussian smoothing filters using intermediate controller based on Finite State Machine simulated in VHDL and prototyped on device technology of XILINX VirtexV FPGA platform. The proposed separable filtering approach is significantly faster and occupied few resources in terms of number of clock cycles and number of multipliers compared to the general two-dimensional convolution implementation. This difference is very important for larger mask sizes because hardware complexity is widely reduced.