engleski

Comparison of PDE-based Sparse Image Inpainting Methods

In recent times image processing and computer vision have become a very attractive fields for applications of partial differential equations (PDEs) and corresponding variational techniques. While applications of PDEs are often associated with denoising problems, there is an increasing number of publications where their potential as inpainting methods is explored. PDEs are frequently used for interpolation problems where most data is given and only filling in the missing data in corrupted areas is needed. It is less frequently used when data is sparser. Goal of the presented paper is to explore the potential of PDEs for a scattered data interpolation problem and importance of selecting good interpolation points for a sparse image or selecting appropriate PDEs if a sparse image is less favourable sparsified. PDE-based interpolation method is described and a number of different PDEs are evaluated on various sparse images.

inpainting, partial differential equation (PDE), diffusion, interpolation, sparse image

nije evidentirano