An Improved Segmentation Technique of Multispectral Image Using Modified Particle Swarm Optimization Algorithm

Murinto, Murinto and prahara, adhi and Heri Ujianto, Erik Iman (2023) An Improved Segmentation Technique of Multispectral Image Using Modified Particle Swarm Optimization Algorithm. [Artikel Dosen]

Text IJASCA.230720.02.pdf - Published Version Download (1MB)

Abstract

The search for optimal solutions to optimization problems often uses meta-heuristic algorithms. This algorithm is generally based on several natural aspects such as biology and physics. One of the advantages of the bioinspired
algorithm is its learning capability to handle optimization problems. This algorithm can be used in multilevel thresholding problems which have recently gained a lot of attention for image segmentation. The problem of
multispectral image segmentation is still challenging and complicated in many applications. Therefore, to overcome this problem, a new multilevel algorithm based on Particle Swarm Optimization (PSO) is proposed in this study. Standard PSO is modified by adding inertial weights and mutations of
position. The experimental results are measured in several parameters, namely computation time (CPU time), Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), Variation of Information (VoI), and Probability Rand Index (PRI). The experimental results show that the proposed PSOt outperforms other
competitive algorithms in terms of stability and convergence rate, which can be applied to practical problems such as multispectral image segmentation.

Item Type:	Artikel Dosen
Subjects:	T Technology > T Technology (General)
Divisi / Prodi:	Faculty of Industrial Technology (Fakultas Teknologi Industri) > S1-Informatics Engineering (S1-Teknik Informatika)
Depositing User:	murinto murinto
Date Deposited:	27 Aug 2025 03:16
Last Modified:	27 Aug 2025 03:16
URI:	http://eprints.uad.ac.id/id/eprint/86669

Actions (login required)

View Item