利用结构约束和样本稀疏表示的图像修复方法

本文探讨了一种利用结构约束和样本稀疏表示，对结构信息缺损较大时的图像修复方法。利用多项式曲线拟合方式修复图像边缘信息，约束结构的修复；采用样本稀疏表示的窄带模型，优先修复结构信息；利用平移块的稀疏表示方法修复纹理信息。仿真结果表明，该方法修复图像质量高，既可较好地修复图像的边缘结构，又能保持结构的整体平滑性。

An Image Inpainting Method by Structural Constraints and Sample Sparse Representation

This paper discusses a way of image inpainting with large defect of structural information by structural constraints and sample sparse representation. Repairing the image edge information by polynomial curve fitting to constrain the structural information; Using narrow-band model of sample sparse representation to repair structural information priority; Completing the texture information by translational block sparse representation method. Simulation results show that proposed method can make higher image quality, and can better repair the structure information, while maintain the structure smoothness integrally.