Difference Curvature Driven Anisotropic Diffusion for Image Denoising Using Laplacian Kernel

Image noise removal forms a significant preliminary step in many machine vision tasks, such as object detection and pattern recognition. The original anisotropic diffusion denoising methods based on partial differential equation often suffer the staircase effect and the loss of edge details when the image contains a high level of noise. Because its controlling function is based on gradient, which is sensitive to noise. To alleviate this drawback, a novel anisotropic diffusion algorithm is proposed. Firstly, we present a new controlling function based on Laplacian kernel, then making use of the local analysis of an image, we propose a difference curvature driven to describe the intensity variations in images. Experimental results on several natural and medical images show that the new method has better performance in the staircase alleviation and details preserving than the other anisotropic diffusions.