Pixel Corrected Adaptive Conductance Function based Diffusion Filter and Image Denoising using Bi-dimensional Empirical Mode Decomposition

In this paper, a method is proposed to denoise image based on semi-adaptive conductance function in anisotropic diffusion filter and bi-dimensional empirical mode decomposition. Here, the color image is utilized in the work to split the image into red, green and blue channels. To each channel component, the local difference value method is implemented where the noise contaminated pixels of the respective channel are replaced with the processed ones in which a Gaussian filter is used for smoothing. The bi-dimensional empirical mode decomposition decomposes the channel into its constituent intrinsic mode functions and the diffusion filter is applied to them. The various function parameters define the extent of diffusion to the channel like connectivity, conductance function, number of iterations and gradient threshold. Here, the conductance function is made semi-adaptive by introducing the gradient value of the image to the threshold parameter along with a preset constant term. The processed intrinsic mode functions, obtained by applying the modified diffusion filter in the conductance function, of each channel are combined and the final image is reconstructed by merging all the three channels. The experimented results thus obtained are evaluated and compared with other existing techniques based on the performance parameters like peak signal-to-noise ratio, mean square error and, structural similarity index and concluded that the proposed method is superior and efficient in both image denoising and feature retention.