Multilevel Thresholding Segmentation based on Otsu’s Method and Autonomous Groups Particle Swarm Optimization for Multispectral Image

Murinto, Murinto and Prahara, Adhi (2019) Multilevel Thresholding Segmentation based on Otsu’s Method and Autonomous Groups Particle Swarm Optimization for Multispectral Image. [Artikel Dosen]

[thumbnail of HASIL CEK_Murinto_Multilevel Thresholding Segmentation based on Otsu’s Method and Autonomous Groups Particle Swarm Optimization for Multispectral Image.pdf]

Text
HASIL CEK_Murinto_Multilevel Thresholding Segmentation based on Otsu’s Method and Autonomous Groups Particle Swarm Optimization for Multispectral Image.pdf
Download (2MB)

Official URL: https://www.ijcaonline.org/

Abstract

Segmentation is a process of division of images into certain
regions based on certain similarities. Multispectral image
consists of several bands with high dimensions, requiring a
different method with the problem of low-dimensional
images. Multilevel thresholding problems based on Otsu
criteria are discussed in this paper. One disadvantage of the
Otsu method is that computing time increases exponentially
according to the number of thresholding dimensions. In this
paper, the Particle Swarm Optimization (PSO) algorithm
combined with the Otsu Method called multilevel
thresholding Autonomous Groups Particles Swarm
Optimization (MAGPSO) is proposed to reduce the two
problems of PSO entrapment in the local minima and the slow
rate of convergence in solving high dimensional problems.
MAGPSO is used for multilevel thresholding image segmentation. The performance of MAGPSO is compared with standard PSO on three natural images. The parameters used to compare the performance of MAGPSO and PSO are the best fitness value, optimal threshold obtained from each
algorithm and the measurement of the quality of segmentation results, namely: SSIM, PSNR, and MSE. From the experimental results show that MAGPSO is better when compared to PSO in image segmentation, in terms of the resulting fitness value and higher SSIM and PNSR values.

Item Type:	Artikel Dosen
Keyword:	Darwinian Particle Swarm Optimization, Hyperspectral Image, Support Vector Machine
Subjects:	Q Science > Q Science (General)
Divisi / Prodi:	Faculty of Industrial Technology (Fakultas Teknologi Industri) > S1-Informatics Engineering (S1-Teknik Informatika)
Depositing User:	murinto murinto
Date Deposited:	18 Apr 2023 09:48
Last Modified:	23 Sep 2023 06:40
URI:	http://eprints.uad.ac.id/id/eprint/43005

Actions (login required)

View Item