Akademik Veri Yönetim Sistemi
VISUAL COMPUTER, cilt.41, sa.13, ss.11025-11051, 2025 (SCI-Expanded, Scopus)
Achieving a fully in-focus image in a scene with depth remains a challenge due to the inherent limitation of camera lenses. Multi-focus image fusion, which combines sharp regions from multiple images captured at different focal distances, offers a practical solution. Block-based fusion methods are simple and efficient, yet determining the optimal block size is crucial. In this study, we present a comprehensive comparison of 16 meta-heuristic optimization algorithms for determining the optimal block size in multi-focus image fusion. Experiments were conducted on five pairs of multi-focus images from the Lytro dataset, and the efficiency of block-based fusion was assessed through both objective evaluations using nine quality metrics and subjective assessments. The results reveal that the Artificial Bee Colony (ABC) algorithm outperformed others in both objective and subjective evaluations. This study not only contributes to the advancement of image fusion techniques but also provides valuable insights into the application of meta-heuristic optimization algorithms in high-dimensional, nonlinear tasks. The code for the application can be accessed at the following GitHub repository: https://github.com/hakbulut60/MultiFocus.