VOL. 11, NO. 24, DECEMBER 2016 ISSN 1819-6608 ARPN Journal of Engineering and Applied Sciences © 2006-2016 Asian Research Publishing Network (ARPN). All rights reserved. www.arpnjournals.com 14075 DESIGN AND MODELING OF LINEAR BACK PROJECTION (LBP) ALGORITHM FOR FIELD PROGRAMMABLE GATE ARRAY (FPGA) Norhidayati Podari 1 , Siti Zarina Mohd Muji 1 , M. Hairol Jabbar 1 and Ruzairi Abdul Rahim 2 1 Faculty of Electrical and Electronic Engineering Universiti Tun Hussein Onn Malaysia, Batu Pahat Johor, Malaysia 2 Faculty of Electrical Engineering Universiti Teknologi Malaysia, Skudai Johor, Malaysia E-Mail: hidayatipodari@gmail.com ABSTRACT This paper focus on designing and modeling of linear back projection (LBP) algorithm for field programmable gate array (FPGA) application. The features provided in FPGA make it the most suitable for embedded system for optical tomography system in data acquisition system. The language supported by FPGA is a Hardware Description Language (HDL). However, manual coding for HDL code spend more time to program. In addition, it increase chances of human error. Therefore, the usage of Matlab Simulink has encouraged many researchers to use it in order to generate HDL code while minimize human error. The LBP algorithm is designed by using Matlab Simulink. From the Matlab Simulink, the HDL code will be generated automatically by using HDL coder which is provided by MathWorks. The HDL code obtained will be downloaded into an FPGA platform of Altera DE2-115. The result obtained shows that the LBP algorithm has been successfully modelled. Therefore, this approach provides an effective method flow for the LBP algorithm to implement in FPGA. Keywords: field programmable gate array, hardware description language, matlab simulink, linear back projection algorithm, HDL coder. INTRODUCTION Linear Back Projection (LBP) algorithm is an algorithm that widely used in optical tomography system. This LBP algorithm is implemented in optical tomography in order to obtain the concentration profile of tomogram image [1]. The concentration profile is obtained from a combination of data projection for each sensor with its computed sensitivity map. LBP algorithm is easy to implement compared to other algorithms. Moreover, it has become popular method among researchers due to low computation, simple and fast respond algorithm [1, 2, 5]. However, it also has disadvantages as well as advantages. The LBP algorithm causes smearing effect when the overlapping image occurs. This is due to the summation of the back projected signals for each pixel [1-4]. The implementation of LBP algorithm in Field Programmable Gate Array (FPGA) is one of the most challenging parts in optical tomography system. Time consuming to develop the project on FPGA causes many researchers did not used it for image reconstruction. Moreover, the difficulty in programming at Register Transfer Level (RTL) also makes the researcher do not used it [6]. Besides that, manual coding of HDL in programming can increase human error [7]. Therefore, Math Works provides an efficient method to program the FPGA. By using HDL coder, user can convert Matlab and Simulink block function to Hardware Description Language (HDL) without having trouble. Here, the purpose of this study is to design and modeling LBP algorithm for FPGA application. Previously, LBP algorithm has been implemented on the microcontroller in optical tomography system. However, it’s contributed many problems such as low capacity memory and slow data processing [8]. Therefore, an FPGA has been proposed to overcome this problem since it has big capacity memory and fast data processing. Altera DE2-115 has been proposed to be used for image reconstruction. OVERVIEW OF THE MODELING DESIGN In this project, LBP algorithm has been implemented in image reconstruction by using Matlab Simulink block. Visual Basic software has been used in order to generate the sensitivity map by developing virtual projection for each transmitter to receiver. The sensor loss value occurs whenever an object blocking the signal transmits by transmitter to receiver. Whilst, the summation of voltage distribution provide information of concentration profile [9]. The tomography image is display based on Equation. (1) where it shows the mathematical formula to reconstruct the image by using LBP algorithm [10]. (1) Where LBP V is voltage distribution, Tx,Rx S is sensor loss value of transmitter and receiver and Tx,Rx M is sensitivity maps of transmitter and receiver. Figure-1 shows the flow to reconstruct tomogram image by using LBP algorithm. Firstly, the sensitivity maps must be processed for every path projection by defining its pixel location. All the elements in the sensitivity maps are summed up. Then, the pixel location of sensitivity map is divided with result obtained from the previous step. The result obtained is multiplied with the sensor loss and m Tx,Rx LBP Tx,Rx Tx=Rx=1 V = S × M 