Double Bilateral Filtering for Image Noise Removal

Bilateral filtering is a popular denoising technique that smooths images while preserving edges by means of a nonlinear combination of adjacent pixel values. We propose a double bilateral filter that extends the classical bilateral filtering for image noise removal. A new median-metric weighting function is introduced by incorporating a median filter into a second bilateral filter. Balancing between two bilateral filters, the double bilateral filter is shown to perform better than or at least equal to the traditional bilateral filter. A variety of images contaminated by various degrees of Gaussian, impulse, and mixed noise were used to assess the performance of this new filtering method. It is indicated that the double bilateral filter outperformed several existing methods in both visual image quality and restored signal quantity.