Image Inpainting Algorithm Based on Fractal Theory
Yuan Da · Dianzi xuebao · 2010
The image inpainting is the hot research points and have widely application.The traditional methods have good results for small area repairing,which can restore the edge structure,but can not handle the texture details and the image with big hole.A new inpainting method based on fractal theory is presented,which uses the total information of image effectively.The relation of fractal dimension and the size range block is discussed,a repair method by multi-scale fractal coding and reconstruction is proposed.In order to strengthen detail information,local fractal iterative is used.Using fractal interpolation,the image is magnified to get good recover quality.Experiments show that the proposed method has good results especially suitable for texture image and large missing image.