Multi-Focus Image Fusion using Unsharp Masking with Discrete Cosine Transform
Manahil Shah, Department of Computer Science, City University of Science and Information Technology, Peshawar, Pakistan.
Sarwar Shah Khan, Department of Computer Science, City University of Science and Information Technology, Peshawar, Pakistan.
Muzammil Khan, Department of Computer and Software Technology, University of Swat, Pakistan.
Salman Ali, Department of Computer Science, City University of Science and Information Technology, Peshawar, Pakistan.
Corresponding Author:
Sarwar Shah Khan (sskhan0092@gmail.com)
Abstract:
Multi-focus image fusion is the process of merging two or more images of the same scene taken with various focus settings into one composite image that includes all of the in-focus areas from each input image. More details and sharpness can be found in the combined image than in any of the single input images. The research proposes a novel approach for fusing images that has two main steps. The initial step is to improve the input images by applying the unsharp masking technique to improve their sharpness and overall quality. This results in more informative images with better contrast and details. The second step involves fusing the enhanced images with the discrete cosine transform (DCT) fusion method, which combines the image information in the frequency domain. This approach aims to produce a resultant image that restrains the most relevant information from the original input images, resulting in improved overall image quality. The novel approach is performed by two datasets such as “clock” and “leaves”. The resultant image is evaluated in two ways qualitative and quantitative metrics. In comparison to baseline approaches, the new method that was tested in an experiment showed effectiveness, enhanced outcomes, and was particularly advantageous for merging images with multiple focuses.
Keywords:
Unsharp Masking; Image Fusion; Multi-Focus; DCT; Transformation Approaches