Removing Smoothing Effects for Color Image Segmentation
Wen-Chang Cheng, Hua-Hung Tseng · 2015
In this study, we described the effects of image smoothing on image segmentation, introduced a method proposed by Yang et al. For removing smoothing effects (hereafter referred as Yang's method), and modified this method to solve image segmentation problems. The results showed that Yang's smoothing method still required further improvements, therefore, two solutions to solving problems related to this method were proposed. In Yang's method, the researchers only considered removing the effects of central pixels in a mask on mask calculation, neglecting the effects of edge pixels in a region of a mask on mask calculation. Therefore, we incorporated an edge mask calculation into Yang's method. Moreover, in Yang's method, distinct color masks are contained in the color components of a single pixel, which cause these components to yield differing adjustment results. Thus, the solution to this problem is to adopt a single-colored mask for color adjustments. The experimental results verified that the method proposed in this study effectively improved Yang's method and removed the effects of smoothing on image segmentation.