الفهرس يوجد فقط 14 صفحة متاحة للعرض العام
 |  | from | 127 |  |  |
| | | from | 127 | | |
المستخلص
The aim of this thesis is to compare between two popular contrast enhancement techniques and automatically select the technique that is most suitable for contrast enhancement of the input images. The included techniques in this study are
Histogram Equalization (HE) and Gray Level Grouping (GLG).
A lot of research work has been done in contrast enhancement but we felt the need of an automatic selector based on each individual input image characteristics. First defined the standard contrast enhancement quality criteria are defined, which are the average distance between pixels on the
gray scale axis and the Tenengrad criterion. Then the techniques HE and GLG are presented in details. A
comparison between them has been performed by applying these techniques on seventy images and mentioning their
advantages and disadvantages. The proposed Fuzzy c-Means classifier as an automatic selector is described. The classifier relies on two features extracted from the input images to take a decision. The features are the average distance between pixels on the gray scale axis and the maximum value of probability intensity level in the image. The decision of the classifier can be: 1) use HE, 2) use GLG, or 3) No clear winner, i.e., both HE and GLG not improve the contrast of the input images.
The proposed method has been tested on thirty images and decisions have been found to mostly agree with the visual
evaluation done by the human eye.