University of Khartoum

Image Processing and Computer Vision: A Comparison between CPU and FPGA

Image Processing and Computer Vision: A Comparison between CPU and FPGA

Show full item record

Title: Image Processing and Computer Vision: A Comparison between CPU and FPGA
Author: Abdelrahman, Hassan Mohammed Warrag
Abstract: Computer vision and image processing are a fundamental issue in computer science. Computer vision is a growing field with many applications in real world. Some of those applications require real time processing. This can be achieved by either improving the core algorithm or by improving the hardware that the algorithm will run on. Regarding the hardware, it is hard to choose between the flexibility provided by general purpose architectures (such as CPUs) and the efficiency provided by purpose built architectures (such as FPGAs). This thesis aims to find a suitable tool for designing image processing and computer vision algorithms in an FPGA and to compare between CPU and FPGA in terms of computational time, resources needed and accuracy( Peak signal to noise ratio). Objectives of this thesis were achieved and a detailed comparison between FPGA and CPU was made. During the research, SIFT algorithm was implemented using MATLAB Simulink and then it was implemented in an FPGA Xilinx Virtex 6. Results using multiple image processing and computer vision algorithms were obtained and it was noticed that in most of the cases, FPGA gave an outstanding performance compared to CPU under specific configurations. A discussion of the reasons behind those results was made and some of the improvements to increase FPGA’s performance in image processing and computer vision applications and future work were suggested
URI: http://khartoumspace.uofk.edu/123456789/25815
Date: 2017


Files in this item

Files Size Format View

This item appears in the following Collection(s)

Show full item record

Share

Search DSpace


Browse

My Account