Image reconstruction based on combination of wavelet decomposition, inpainting and texture synthesis

Abstract

Digital inpainting provides a means for reconstruction of damaged portions of an image. Although the inpainting basics are straightforward, most inpainting techniques published in the literature are only suitable for remarkable small portion or smooth color image. In order to avoid such shortcomings, we present a new algorithm for digital reconstruction based on combination of wavelet decomposition, Surface-based/PDE-based inpainting and texture synthesis. In which, wavelet transform at first decomposes the image into high frequency and low frequency level parts. Subsequently, CSRBF which is generally used for surface interpolation or PDE-based inpainting is employed for low frequency level and texture synthesis is used for high frequency level. It results in that not only slightly portion but also the common blotched image can be reconstructed with high quality. Especially, our algorithm makes many difficult cases for other methods possible.