A Study on Implementation of Image Processing For Multiplication Image Processing Application: Optimized implementations of multiplication for image and signal processing applications using EPGAs

Abstract

Multiplication is the kernel operation used in many imageand signal processing applications. In this paper, we present the design and(EPGA) implementation of multiplier architectures for use in image and signalprocessing applications. The designs are optimized for speed which is the mainrequirement in these applications. First design involves computation of densemultiplication which is used in image processing application. The design hasbeen implemented on Virtex-4 EPGA and the performance is evaluated by computingthe execution time on EPGA. Implementation results demonstrate that it can providea throughput of 16970 frames per second which is quite adequate for most imageprocessing applications. The second design involves multiplication oftri-matrix (three matrices) which is used in signal processing application. Theproposed design for the multiplication of three matrices has been implementedon Spartan-3 and Virtex-II Pro platform EPGAs respectively. Implementationresults are presented which demonstrate the suitability of EPGAs for suchapplications.