Search

The concept of S-rough set is introduced to represent images as S-rough sets for fitting complex background and noisy environment by combining migration characteristics. This model can remove pixels with singular nature from original object dynamically. A more adaptable technique of image thresholding segmentation is proposed by establishing a rough entropy to compromise the values of object roughness and background roughness. In order to adjust the transfer of singular points and avoid choice of granule size, images are further described by variable precision S-rough sets in combination with the notion of inclusion degree to extract object from background. Experimental results show that the proposals have better performance of thresholding segmentation of noised images. [ABSTRACT FROM AUTHOR]