An Image Inpainting Algorithm Based on Local Geometric Similarity

Authors:
Pan Qi;Xiaonan Luo;Jiwu Zhu
Affiliations:
Computer Application Institute, Sun Yat-Sen University, Guangzhou, China and Key Laboratory of Digital Life(Sun Yat-Sen University), Ministry of Education,;Computer Application Institute, Sun Yat-Sen University, Guangzhou, China and Key Laboratory of Digital Life(Sun Yat-Sen University), Ministry of Education,;Dept. Marketing, Guangdong Pharmaceutical University, Guangzhou, China
Venue:
MMM '09 Proceedings of the 15th International Multimedia Modeling Conference on Advances in Multimedia Modeling
Year:
2009

Citing 8
Cited 0

Image inpainting

Proceedings of the 27th annual conference on Computer graphics and interactive techniques
Synthesizing natural textures

I3D '01 Proceedings of the 2001 symposium on Interactive 3D graphics
An $\cal O(N)$ Level Set Method for Eikonal Equations

SIAM Journal on Scientific Computing
Vector-Valued Image Regularization with PDEs: A Common Framework for Different Applications

IEEE Transactions on Pattern Analysis and Machine Intelligence
Edge-directed prediction for lossless compression of natural images

IEEE Transactions on Image Processing
New edge-directed interpolation

IEEE Transactions on Image Processing
Inpainting of Binary Images Using the Cahn–Hilliard Equation

IEEE Transactions on Image Processing
Novel sequential error-concealment techniques using orientation adaptive interpolation

IEEE Transactions on Circuits and Systems for Video Technology

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper proposes a novel noniterative orientation adaptive image inpainting algorithm. Assuming the image can be locally modeled, the filling process is formulated as a linear optimization problem, which the optimal coefficients can be adapted to match an arbitrary-oriented edge based on local geometric similarity . We provided A Weighted Least Square (WLS) method is provided to offer a convenient way of finding the optimal solution, which the weight function is selected based on the non local means. We also present Group Marching method (GMM) as the propagation scheme such that sharp edges are well propagated into the missing region layer by layer while maintaining the local geometric similarity. A number of examples on real and synthetic images demonstrate the effectiveness of our algorithm.