Image Denoising Using Spatial Gradient Based Bilateral Filter and Minimum Mean Square Error Filtering
Mayank Tiwari, Bhupendra Gupta · Procedia Computer Science · 2015
In this paper we propose noise removal approach using spatial gradient based bilateral filter and minimum mean square error filtering. The proposed method consist two-steps process. In first step, we generate a reference image from the given noisy image, by extracting (2W + 1) × (2W + 1) size patches from it and then apply proposed spatial gradient based bilateral filter on each of these patches. To reduce the mean square error, in second step we apply minimum mean square error (MSE) filter on the reference image. Generally all the noise removal approaches change the natural appearance of the restored image, while the proposed method restores the image without affecting its natural appearance.