Memory Based Bees Algorithm With Lévy -Flights For Multilevel Image Thresholding

Nahlah Mohammad Amin Shatnawi (P57799)

Please use this identifier to cite or link to this item: https://ptsldigital.ukm.my/jspui/handle/123456789/513497

Full metadata record

DC Field	Value	Language
dc.contributor.advisor	Shahnorbanun Shahran, Associate Dr.	-
dc.contributor.author	Nahlah Mohammad Amin Shatnawi (P57799)	-
dc.date.accessioned	2023-10-16T04:37:18Z	-
dc.date.available	2023-10-16T04:37:18Z	-
dc.date.issued	2013-11-05	-
dc.identifier.other	ukmvital:75114	-
dc.identifier.uri	https://ptsldigital.ukm.my/jspui/handle/123456789/513497	-
dc.description	Thresholding is a method for region-based image segmentation, which is important in image processing applications such as object recognition Multilevel thresholding used to find multiple thresholds values. Most of the multilevel thresholding methods use PSNR (Peak Signal to Noise Ratio) as final evaluation criteria; however only few works use PSNR to find suitable threshold values. In order to find the values, usually the existing multilevel thresholding algorithms are trapped in exhaustive search computation time problems. To solve this problem many optimization algorithms have been adopted, such as the Bees Algorithm (BA) that had been found its usefulness in image processing. However, the current BA has two disadvantages which are: Firstly, it does not fully imitate all physical and social aspect of bees' nature such that it doesn't up to its potential. Secondly, it requires a quite number of tuneable and randomized parameters. The objectives of this research are to propose: 1) a BA based multilevel image thresholding technique that uses PSNR as method to select number of threshold values that will solve the exhaustive search computation time issue. 2) a memory based BA (MBA) that will improve the algorithm generally by copying decision making capabilities of bees; 3) a Lévy flight movement based on the MBA (LMBA) to reduce the algorithm randomized and tuneable parameters; 4) a multilevel image thresholding algorithm based on LMBA with PSNR for segmentation of various types of images. In the MBA, two kinds of bee memory are proposed, local and global, which are meant for memory of individual bee and the colony respectively. In the LMBA, the normal random initialization and bee movements are replaced by random walk of Lévy-flight. Lévyflight uses Lévy probability distribution that represents an optimal foraging strategy in many foraging patterns in nature including the honey bees. Local-MBA, global-MBA, MBA and LMBA are tested using several benchmark functions. They had obtained approximately 59.34%, 73.02%, 74.9% and 81.13% improvement on mean number of evaluations over the basic BA respectively. At the end, the BA, MBA and LMBA are tested to binarize various types of images using multilevel thresholding algorithm based on the PSNR quality index. In the statistical test, the LMBA outperforms BA and MBA in both computation time and quality. The proposed methods have great potential to be used in other optimization problems, and for many multi-level image thresholding applications like the License Plate Recognition (LPR) and medical imaging.,PhD	-
dc.language.iso	eng	-
dc.publisher	UKM, Bangi	-
dc.relation	Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat	-
dc.rights	UKM	-
dc.subject	Memory Based Bees Algorithm With Lévy	-
dc.subject	Flights For Multilevel Image Thresholding	-
dc.subject	Image segmentation	-
dc.title	Memory Based Bees Algorithm With Lévy -Flights For Multilevel Image Thresholding	-
dc.type	Theses	-
dc.format.pages	196	-
dc.identifier.callno	TA1638.4 .N334 2013 3	-
dc.identifier.barcode	000565	-
Appears in Collections:	Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat

Files in This Item:

File	Description	Size	Format
ukmvital_75114+Source01+Source010.PDF Restricted Access		7.68 MB	Adobe PDF	View/Open

Show simple item record Recommend this item