Please use this identifier to cite or link to this item:
|Title:||Conditional Toggle Mappings: Principles and Applications|
|Authors:||VELASCO-FORERO Santiago; ANGULO Jesus; SOILLE Pierre|
|Citation:||JOURNAL OF MATHEMATICAL IMAGING AND VISION vol. 48 no. 3 p. 544–565|
|Type:||Articles in periodicals and books|
|Abstract:||We study a class of mathematical morphology filters to operate conditionally according to a set of pixels marked by a binary mask. The main contribution of this paper is to provide a general framework for several applications including edge enhancement and image denoising, when it is affected by salt-and-pepper noise. Specifically, we proposed a novel basic pair of morphological operators (dilation/erosion) by constraining the noisy contaminated pixels of the image. We demonstrate the these operators can be incorporated into a conditional toggle mapping to produce a modification of classical shock filters by including the information contained in a seed mask. Applications in edge enhancement, morphological segmentation and image denoising are explored. In the case of edge enhancement, conditional toggle mapping allows to avoid unpleasant effect in morphological regularisation for segmentation, as classical shock filters, but it allows a fast computation thanks to the small number of iterations necessary to convergence. Additionally, we compare our approach with more advanced complex techniques, such as NL-medians and kernel regression for the case of denoising in salt-and-pepper corrupted images. Experimental comparison show that the proposed algorithm can provide better performance for suppressing impulse noise with high noise level (>60%) than many other well-known techniques.% Our approach behaves better for very high noise level high (> 75%$, yielding a performance which is relatively stable even for extremely high levels (> 90%).|
|JRC Directorate:||Space, Security and Migration|
Files in This Item:
There are no files associated with this item.
Items in repository are protected by copyright, with all rights reserved, unless otherwise indicated.