Regulating The Degree Of Contrast Enhancement In Global Histogram Equalization-Based Method For Grayscale Photo Processing
Global Histogram equalization (GHE) is a popular image contrast enhancement method. However, it is rarely used on photo processing because it tends to create noise-artifacts, especially in simple-structure-image. A few GHE-based methods have been proposed to address this issue but whether they are n...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English |
| Published: |
2007
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/5310/1/FK_2007_75a.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Global Histogram equalization (GHE) is a popular image contrast enhancement method. However, it is rarely used on photo processing because it tends to create noise-artifacts, especially in simple-structure-image. A few GHE-based methods have been proposed to address this issue but whether they are noise-artifacts-proof remains questionable. This is because the methods are fully automatic and the evaluation conducted was not comprehensive.
A novel automatic GHE-based method called Minimum Mean Brightness Error Bi-Histogram Equalization (MMBEBHE) has been proposed in this thesis. It has been evaluated thoroughly together with the existing automatic methods. The results have proven that none of the automatic GHE-based methods is noise-artifacts-proof. The conclusion has motivated author to look into scalable GHE-based methods that allows user to regulate the degree of contrast enhancement. A novel scalable GHE-based method called Recursive Mean-Separate Histogram Equalization (RMSHE) has been proposed in this thesis. It has been evaluated thoroughly together with other two existing scalable methods - Clip Limited Adaptive HE (CLAHE) and Stark’s Adaptive HE (StarkAHE). The results of separate evaluations consistently showed that none of the three methods could effectively enhance the contrast of simple-structure-image without creating any noise-artifacts. Another novel scalable GHE-based method called Scalable Global Histogram Equalization with Selective Enhancement (SGHESE) has been developed then to overcome the limitation of the existing methods. Evaluation results showed that SGHESE could enhance the image’s contrast effectively without creating any noise-artifacts. The results of subjective evaluation involving human observer also showed that the preference level of SGHESE was significantly higher compared to those of other methods. Finally, the thesis recommends extending the study of SGHESE to color image processing because majority of the images nowadays are color images. |
|---|