Performance analysis of segmentation using SSR under different noise conditions
Sumit Kumar, Ayush Jha · 2017
In this paper, we propose image segmentation by Suprathreshold Stochastic Resonance (SSR) after filtering with anisotropic diffusion (AD). AD, which is used to remove the noise and to preserve the significant detail of image like lines, edges etc. It is followed by SSR, which uses the noise for segmentation of the noisy and blurred color images having different brightness values. This algorithm has many significant improvements over some art of methods. This technique excels over other traditional segmentation techniques in terms of correlation coefficient, change in object positions and number of mismatch pixels as performance parameters. We validate this algorithm using Gaussian, Uniform, Laplacian, Poisson and Gamma noise.