International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 12 | Dec 2025
p-ISSN: 2395-0072
www.irjet.net
Multi-Objective Filter Optimization Using Grey Wolf Optimizer for Medical Image Enhancement: A Comparative Study with Grey-Level Contrast Enhancement Techniques Monika1, Dr. Chandan Kumar2 1Research Scholar, Department of Computer Science and Engineering, Career Point University, H.P., India
2Associate Professor, Department of Computer Science and Engineering, Career Point University, H.P., India
-----------------------------------------------------------------------***--------------------------------------------------------------------
Abstract - Medical image enhancement is critical for
Traditional image enhancement approaches, such as histogram equalization and contrast stretching, operate on fixed parameters and single-objective functions. This limitation creates inherent trade-offs: while aggressive enhancement may improve contrast, it often amplifies noise or produces unnatural appearances that reduce clinical acceptability. Contemporary medical imaging requires enhancement methods that (1) intelligently balance competing objectives, (2) adapt to local image characteristics through dynamic thresholding, and (3) preserve clinically relevant details while suppressing artifacts [3].
improving diagnostic accuracy and clinical outcomes. This research proposes a novel approach to optimize image enhancement filters through multi-objective optimization using the Grey Wolf Optimizer (GWO) with dynamic threshold adaptation. The study focuses on two primary objectives: (1) improving filter performance by optimizing multiple goals simultaneously for dynamic threshold adaptation, and (2) comprehensively comparing the proposed enhancement methods against existing gray-level contrast enhancement techniques using various performance metrics. The proposed framework integrates Convolutional Neural Networks (CNNs) with metaheuristic optimization and Non-Local Means (NLM) filtering to achieve superior image quality while preserving anatomical details. Experimental evaluation on ultrasound, CT, and MRI datasets demonstrates that the proposed multiobjective optimization approach outperforms traditional methods such as Histogram Equalization (HE), ContrastLimited Adaptive Histogram Equalization (CLAHE), and linear contrast stretching across PSNR, SSIM, UQI, and NCC metrics. The results highlight the effectiveness of dynamic threshold mechanisms in adapting to local image characteristics, thereby enabling robust enhancement across diverse imaging modalities and noise conditions.
The advent of metaheuristic optimization algorithms, particularly nature-inspired techniques like the Grey Wolf Optimizer (GWO), offers promising solutions for multiobjective parameter tuning. GWO has demonstrated superior performance compared to other optimization methods in high-dimensional, nonlinear, multi-modal problems inherent in image processing [4]. By incorporating GWO with spatial filtering techniques (such as convolutional filtering and NonLocal Means filtering) and dynamic threshold mechanisms, we can develop an adaptive enhancement framework that optimizes image quality across multiple dimensions simultaneously [5]. 1.1 Research Problem
Key Words: Image enhancement, multi-objective optimization, Grey Wolf Optimizer, dynamic thresholding, medical imaging, comparative evaluation.
Medical imaging modalities such as Magnetic Resonance Imaging (MRI), Computed Tomography (CT), ultrasound, and X-ray radiography are essential diagnostic tools in contemporary clinical practice. However, these images are frequently degraded by low contrast, noise, blur, and artifacts that reduce image interpretability and diagnostic confidence [1]. The quality of acquired images depends on multiple factors including imaging hardware capabilities, acquisition parameters, patient motion, and environmental noise. These limitations necessitate sophisticated image enhancement techniques capable of simultaneously optimizing multiple quality aspects without introducing artificial artifacts [2].
1. INTRODUCTION Medical imaging modalities such as Magnetic Resonance Imaging (MRI), Computed Tomography (CT), ultrasound, and X-ray radiography are essential diagnostic tools in contemporary clinical practice. However, these images are frequently degraded by low contrast, noise, blur, and artifacts that reduce image interpretability and diagnostic confidence [1]. The quality of acquired images depends on multiple factors including imaging hardware capabilities, acquisition parameters, patient motion, and environmental noise. These limitations necessitate sophisticated image enhancement techniques capable of simultaneously optimizing multiple quality aspects without introducing artificial artifacts[2].
© 2025, IRJET
|
Impact Factor value: 8.315
Traditional image enhancement approaches, such as histogram equalization and contrast stretching, operate on fixed parameters and single-objective functions. This limitation creates inherent trade-offs: while aggressive enhancement may improve contrast, it often amplifies noise
|
ISO 9001:2008 Certified Journal
|
Page 800