Criminisi-Based Sparse Representation for Image Inpainting
Authors:
- Gaolong HuGHGaolong Hu
- Ling XiongLXLing Xiong
- 1 April 2017
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 389-393
Abstract
Image inpainting is an important area in image processing, which has important application significance in the fields of cultural relic protection, inpainting of damaged old photographs and removal of redundant objects on images. Aiming at the drawbacks of the best matching block search and fill in Criminisi algorithm, and the superior performance of sparse representation in signal recovery, a Criminisi algorithm combined with sparse representation is proposed in this paper. In the proposed algorithm, the sparse representation inpainting method is used to replace the best matching patch search in Criminisi algorithm, and the marked areas to be inpainting are optimized and the priority of credibility is improved. The experimental results of image inpainting show that the proposed algorithm has strong adaptability and achieved good inpainting effect.Keywords
This publication has 4 references indexed in Scilit:
- Image Inpainting : Overview and Recent AdvancesIEEE Signal Processing Magazine, 2013
- For most large underdetermined systems of equations, the minimal 𝓁1‐norm near‐solution approximates the sparsest near‐solutionCommunications on Pure and Applied Mathematics, 2006
- Vector-valued image regularization with PDEs: a common framework for different applicationsIEEE Transactions on Pattern Analysis and Machine Intelligence, 2005
- Region Filling and Object Removal by Exemplar-Based Image InpaintingIEEE Transactions on Image Processing, 2004