Anisotropic diffusion filter using Haar wavelet

In this paper, an adaptive anisotropic diffusion filter with Haar wavelet transform was proposed. Firstly, Haar transform of noisy image was performed. Then the centers of gravity of histograms of wavelet coefficients were used as a conductance parameter, K in anisotropic diffusion equations. In this context, an adaptive filter where user intervention is not required has been obtained. Finally, the approach was tested with well known test images with speckle noise. It was observed that edges in image were preserved while smoothing within regions.