الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 133 |  |  |
| | | from | 133 | | |
Abstract
Multiple imaging modalities can complement each other to provide more information to understand the real worlds of objects than the use of a single modality. Image fusion aims to generate a single image which contains more precise reliable visualization of the objects than any of the source images. Such a fused image should provide extended information and better perception for human vision or computerized vision tasks. All source images need to be accurately aligned or spatially registered before fusion. Image fusion has been investigated by many researchers in various fields. Several great works in the last decade established the basic principles and sub-specialties evolved and grew. Recent efforts have led to the development of a number of algorithms, performance assessment, processing approaches and promising applications.
This thesis introduces an overview study about image fusion techniques. These techniques are implemented and applied to a set of famous applications. Six different image fusion techniques are implemented, analyzed, and compared. These techniques are: the Laplacian Pyramid, the Wavelet Transform, the Computationally Efficient Multi-scale Image Fusion (CEMIF) method, the Multi-focus Technique based on Spatial Frequency, the Curvelet Transform, and the Contourlet Transform. Two methods for fusing color images are implemented, analyzed, and compared. These methods are: fusion in RGB model and fusion in YIQ model.
Fusion of multi-focus images application and fusion of multi-exposure images application are introduced including fusion of 2D images and fusion of color images. The six image fusion techniques are implemented and compared using four measures of performance; the entropy, the spatial frequency, the mutual information, and amount of edge information.
In all fusion applications, the fused image has better eye perception than both source images. It has been shown that no significant difference between the fusion techniques. The best fusion algorithm and the best performance measure are application dependent i.e. the best fusion criterion and the best performance measure should be linked with the specific application.