A comprehensive survey of image segmentation: clustering methods, performance parameters, and benchmark datasets.

Area of Science:

• Computer Vision
• Image Processing

Background:

• Image segmentation is crucial for extracting information from images, with applications ranging from object detection to medical image analysis.
• Clustering is a common approach for image segmentation due to unlabelled image pixels.
• This paper reviews clustering-based image segmentation methods.

Purpose of the Study:

• To provide a comprehensive review of existing clustering-based image segmentation methods.
• To focus on partitional clustering methods due to their computational advantages.
• To discuss performance evaluation parameters and benchmark datasets.

Main Methods:

• Survey of hierarchical and partitional clustering methods for image segmentation.
• Detailed examination of partitional clustering categories: K-means, histogram-based, and meta-heuristic methods.
• Review of quantitative performance evaluation parameters for segmentation results.

Main Results:

• Partitional clustering methods are computationally more efficient than hierarchical methods.
• Partitional methods are further categorized into K-means, histogram-based, and meta-heuristic approaches.
• A review of performance metrics and benchmark datasets for segmentation evaluation is presented.

Conclusions:

• Clustering, particularly partitional methods, is a vital technique for image segmentation.
• Understanding different clustering approaches, performance metrics, and datasets is essential for effective image segmentation.
• This review serves as a guide to the landscape of clustering-based image segmentation.