@Article{dedt.2025.066344,
AUTHOR = {Muhammad Osama, Sarwar Shah Khan, Sajid Khan, Muzammil Khan, Mian Muhammad Danyal, Reshma Khan},
TITLE = {A Systematic Comparison of Discrete Cosine Transform-Based Approaches for Multi-Focus Image Fusion},
JOURNAL = {Digital Engineering and Digital Twin},
VOLUME = {3},
YEAR = {2025},
NUMBER = {1},
PAGES = {17--34},
URL = {http://www.techscience.com/dedt/v3n1/63442},
ISSN = {},
ABSTRACT = {Image fusion is a technique used to combine essential information from two or more source images into a single, more informative output image. The resulting fused image contains more meaningful details than any individual source image. This study focuses on multi-focus image fusion, a crucial area in image processing. Due to the limited depth of field of optical lenses, it is often challenging to capture an image where all areas are in focus simultaneously. As a result, multi-focus image fusion plays a key role in integrating and extracting the necessary details from different focal regions. This research presents a comparative analysis of various Discrete Cosine Transform (DCT)-based methods for multi-focus image fusion. The primary objective is to provide a clear understanding of how these techniques differ based on mathematical formulations and to compare their visual and statistical performance. The analysed methods include: DCT + Variance and DCT + Variance + Consistency Verification (CV), DCT + Correlation Coefficient (CC) and DCT + CC + CV, DCT + Singular Value Decomposition (SVD) and DCT + SVD + CV, DCT + Sharpening and DCT + Sharpening + CV, DCT + Correlation Energy (Corr_Eng) and DCT + Corr_Eng + CV, Through experimental evaluations, the study finds that DCT + Variance and DCT + Variance + CV consistently deliver superior results across all tested image sets. The performance of these methods is evaluated using six different quantitative metrics, demonstrating their effectiveness in enhancing image quality through fusion.},
DOI = {10.32604/dedt.2025.066344}
}