UniPolar impulse noise removing from digital images based on estimation of noises spatial distribution
This article introduces a method to detect and remove unipolar impulse noises from digital images. In this study, proposed method is a repetitive procedure which each repeat contain three steps. Also, proposed method can be based on two detection algorithms that are introduced to detect unipolar imp...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | fas |
| Published: |
Semnan University
2015-01-01
|
| Series: | مجله مدل سازی در مهندسی |
| Subjects: | |
| Online Access: | https://modelling.semnan.ac.ir/article_1688_04cfc1a54150fa3acae42a58a21a4ba4.pdf |
| Summary: | This article introduces a method to detect and remove unipolar impulse noises from digital images. In this study, proposed method is a repetitive procedure which each repeat contain three steps. Also, proposed method can be based on two detection algorithms that are introduced to detect unipolar impulse noises. In this research, unipolar impulse noises with 3 different distribution types are modeled and added to original image to produce noised image. Then, image is reconstructed based on proposed method from produced noised image. Finally, reconstructed image is evaluated in comparison to original image. Detected noises by introduced algorithms are evaluated by Overall Accuracy index. Also, quality of the reconstructed image is evaluated by Peak Signal to Noise Ratio (PSNR) index. Evaluation results are shown that proposed method based on two detection algorithms is better than regular and adaptive two-pass median filters. Quality of the reconstructed images by proposed method is high than results of regular and adaptive two-pass median filters. The average enhancement in quality is equivalent with increasement 5.37 and 4.17 decibels in PSNR index respectively. |
|---|---|
| ISSN: | 2008-4854 2783-2538 |