Intelligent Engineering Systems through Artificial Neural Networks, Volume 16
71 A Novel Approach to Reduce Universal Noise in Gray Images Using Fuzzy Filters
Download citation file:
- Ris (Zotero)
- Reference Manager
This paper presents the fuzzy filter consisting of the two filters, Gaussian and Impulse noise filters for eliminating the mixed noise. For Gaussian filter, the fuzzy set called “small” is derived to represent the disorder in a pixel arising out of neighborhood corrupted with Gaussian. Next, expression for correction is developed based on the intensity of the central pixel and the membership function. Similarly, the correction for the Impulse noise is developed by finding the middle ranking pixels in the neighborhood of the central pixel. The difference between the average of the middle ranking pixels and the central pixels is used to get the membership function which when multiplied by the difference gives the correction. Next, the presence of noise is detected by finding the aggregate of four highest memberships of the neighborhood pixels. If this aggregate is more than the threshold then there is Gaussian noise otherwise impulse noise. Accordingly, the corrupted pixel will be corrected by the correction term. The results are found to be satisfactory.