Dissertation/Thesis Abstract

Image scaling with three different interpolation techniques and their comparisons
by Patra, Pooja, M.S., California State University, Long Beach, 2009, 72; 1481630
Abstract (Summary)

The objective of this study is to demonstrate image upscaling and to recreate the scaled image with different interpolation algorithms. We prove with our implementation of Catmull-Rom interpolation algorithm that is better than the existing nearest neighbor and linear interpolation algorithm. We compare the results obtained with three different interpolation techniques and prove why Catmull-Rom based method that we have implemented, produces the best image quality as compared to the other two methods.

The different interpolation algorithms are explained with in-depth explanation of nearest neighbor, linear interpolation and then finally Catmull-Rom based interpolation which is a bicubic interpolant and better than the other two methods proved by comparison of image qualities of output images. Horizontal scaling is performed followed by vertical scaling to a factor of 1.25. The scaled image is then recreated with different interpolation algorithms. The drawback with this implementation is that it only accepts .ppm images as input and the output image is in .ppm format too. However, softwares like IrfanView can easily convert from .ppm to any other format. The algorithms have been implemented in C language.

Indexing (document details)
Advisor: Englert, Bukhard
Commitee:
School: California State University, Long Beach
School Location: United States -- California
Source: MAI 48/04M, Masters Abstracts International
Source Type: DISSERTATION
Subjects: Computer science
Keywords:
Publication Number: 1481630
ISBN: 9781109656985
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest