Image De-noising Based On Adaptive Kernel Regression

Image de-noising is a key aspect of the image processing,in practical applications, it is often as the pre-processing of the image processing and recognition,it is the basic of the follow-up processing,such as image segmentation and image recognition,etc.In order to eliminate or attenuation exists in the image of the noise,at the same time as much as possible to retain the image details,a novel image de-noising algorithm based on adaptive kernel regression is proposed in this paper.Here,the image can be regarded as a surface of the image intensity function and a second order Taylor polynomial surface to approximate the original,the original model can adaptive the vary of the local area near the center pixel,adjust the shape of the kernel function windows,to minimize the approximation error.The final studies show that our de-noising result is well,with high Edge-preserving quality.