Novel Image Inpainting Algorithm Based on CSRBF Interpolation
Wang Wei-hong①②; Qin Xu-jia①③
①College of Software Engineering, Zhejiang University of Technology, Hangzhou 310032, China;②State Key Lab. of Software Development Environment, Beijing University of Aeronautics and Astronautics,
Beijing 100083, China;③State Key Lab of CAD&CG, Zhejiang University, Hangzhou 310027,China
Abstract：A novel algorithm for image inpainting based on Compactly Supported Radial Basis Functions(CSRBF) interpolation is proposed. The algorithm looks an image as an incomplete 3D points set. The 2D image inpainting problem is converted into implicit surface reconstruction problem from 3D points set. To construct the implicit surface for approximating the points set by using Radial Basis Functions(RBF) can interpolate the points in absent portion, and then resample from the constructed surface can calculate the pixels’ value of damaged or removed portion on the image. Using CSRBF, the matrix of corresponding system of the linear algebraic equations is spare and bounded. So it can decrease the complexity of RBF algorithm. A number of examples on image inpainting demonstrates that the algorithm is an practical one and using the algorithm can obtain good results.
2006, Vol. 28 
