Enhancement of Single and Composite Images Based on Contourlet Transform Approach

Image enhancement is an imperative step in almost every image processing algorithms.
Numerous image enhancement algorithms have been developed for gray scale images
despite their absence in many applications lately. This thesis proposes hew image
enhancement techniques of 8-bit single and composite digital color images. Recently, it
has become evident that wavelet transforms are not necessarily best suited for images.
Therefore, the enhancement approaches are based on a new 'true' two-dimensional
transform called contourlet transform. The proposed enhancement techniques discussed
in this thesis are developed based on the understanding of the working mechanisms of the
new multiresolution property of contourlet transform. This research also investigates the
effects of using different color space representations for color image enhancement
applications. Based on this investigation an optimal color space is selected for both single
image and composite image enhancement approaches. The objective evaluation steps
show that the new method of enhancement not only superior to the commonly used
transformation method (e.g. wavelet transform) but also to various spatial models (e.g.
histogram equalizations). The results found are encouraging and the enhancement
algorithms have proved to be more robust and reliable.