Patch based image restoration using adaptive bilateral filtering

This paper proposes noise suppression using adaptive bilateral filtering especially for Gaussian and speckle noise. This filter is used for the sharpness enhancement of degraded images and performs well for Gaussian noise and speckle noise reduction. Bilateral filtering is a non-linear technique which can smudge the image while preserving strong edges. It is a noniterative image denoising method which has the capability to crumble an image into diverse scales devoid of causing haloes. In order to level the noisy image, it utilizes the gradient value of the noisy image. Bilateral filtering is implemented with varying sigma values. Noisy image is splitted into patches and the sigma values are calculated for implementing adaptive bilateral filtering. The image denoising performance is further improved by adaptive bilateral filtering with less computational complexity. An experimental result of the proposed image denoising algorithm achieves better performance compared to other filtering methods. Quality metrics like PSNR and MSE provide better result when compared to other methods.