Multicore algorithms for image alignment

Ward, Tristan (2011) Multicore algorithms for image alignment. [USQ Project]

[img] PDF
Ward_2011.pdf

Download (1906Kb)

Abstract

Parallel processing is an emerging trend in modern computing. Traditional software development paradigms often forsake parallelism in their approach to produce
algorithms. Applications developed then effectively relinquish any potential performance benefits gained by using multi–core processing hardware that is
presently available. The fundamental idea of using parallel processing is applied to medical research and the results are reported in this dissertation. Advancements in
technology within this field have the potential to greatly streamline processing, thereby directing scientific attention back to research.

Advances in medical microscopy are presently being hindered by the substantial time involved with the construction of panoramic imagery. The predominate purpose and focus of the project is to investigate and develop the automation of
image alignment and noise reduction to a series of microscopy photographs, using the performance advantages of multiple processor cores. The output of the algorithms is the formation of a single microscopic panoramic image.

Pursuing the intention of parallelism, the implementation involves adaptation of certain recognized algorithms. Alignment of the images is achieved by correlation, a
typical form of digital signal processing technique to measure the similarity between images. Reducing the noise in the photographs is accomplished by a computationally efficient median filter. The algorithms which were evaluated
provide a means of automated batch image construction without the need for user intervention. This has the potential to save time by multi–threading on as many
processor cores as possible for the system that it executes on.


Statistics for USQ ePrint 22709
Statistics for this ePrint Item
Item Type: USQ Project
Refereed: No
Item Status: Live Archive
Faculty / Department / School: Historic - Faculty of Engineering and Surveying - Department of Electrical, Electronic and Computer Engineering
Supervisors: Leis, John
Date Deposited: 14 Jan 2013 02:54
Last Modified: 16 Jan 2013 07:17
Uncontrolled Keywords: multicore algorithm, image alignment
Fields of Research : 09 Engineering > 0906 Electrical and Electronic Engineering > 090609 Signal Processing
URI: http://eprints.usq.edu.au/id/eprint/22709

Actions (login required)

View Item Archive Repository Staff Only