Date of Award

Summer 2021

Publication Type

Thesis

Degree Name

M.A.Sc.

Department

Electrical and Computer Engineering

First Advisor

A. Ahmadi

Second Advisor

M. Khalid

Third Advisor

H. Wu

Keywords

Image processing, Image enhancement algorithm, Power consumption

Rights

info:eu-repo/semantics/openAccess

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Abstract

Image processing plays very crucial role in this digital human world and has rapidly evolved with the development of computers, mathematics and the real-life demand of variety of applications in wide range of areas. This wide range of areas includes remote sensing, machine/ robot vision, pattern recognition, medical diagnosis, video processing, military, agriculture, television, etc. Image processing has two important components which are image enhancement and information extraction. Since image enhancement works at the front end with the initial raw inputs, it works like a backbone in image processing. When it comes to implementing these image enhancement techniques and developing applications, these tasks are bit demanding in the choice of processing units because the demand of high resolution. This emerges the necessity of a high speed, powerful and cost-effective processing unit. In this thesis we present an improved image enhancement algorithm in terms of performance and its implementation on FPGA as they satiates the necessity of high speed, powerful and cost-effective processing unit by providing flexibility, parallelization, pipelining and reconfigurability. We have performed a high level synthesis by using MATLAB and implemented an improved image enhancement algorithm on Cyclone V by using Quartus Prime. We have considered an X-ray image size of 1000x1920p for implementation and achieved decent PSNR values and hardware resource utilization along with the better visual interpretability by our proposed improvements. For achieving a better execution time and power consumption we also offer the task parallelism for the algorithm.

Share

COinS