A Combined Filtering Approach for Image Denoising

The quality of digital images are degraded by the noises. In this paper, a combined filtering approach is proposed to remove the salt & pepper noise. It consist of three value weighted filter and traditional median filter. In three value weighted filter, a variable local window is applied to find the noisy pixels. By using the noise free pixels in that window, the noisy pixels are reconstructed using three value method. In median filter, a variable local window is applied to reconstruct the noisy pixels. In that window the central noisy pixel is reconstructed by the median value of the pixels in that window. At last wavelet transform is used to increase the resolution of the reconstructed image. The experimental results shows that the proposed filter denoises the image and improves the resolution when compared to the existing methods.