Previous [ 1] [ 2] [ 3] [ 4] [ 5] [ 6] [ 7] [ 8] [ 9] [ 10]


Journal of Information Science and Engineering, Vol. 28 No. 4, pp. 755-770 (July 2012)

An Effective Content-Aware Image Inpainting Method*

Department of Computer Science and Engineering
National Chung Hsing University
Taichung, 402 Taiwan

Image inpainting is a technique for restoring damaged old photographs and removing undesired objects from an image. The basic idea behind the technique is to automatically fill in lost or broken parts of an image using information from the surrounding area. The challenge of current inpainting algorithms is to restore both texture and structure characteristic information for large and thick damaged regions. This paper presents a new hybrid image inpainting method based on Bezier curves which combines the exemplar-based inpainting technique and the edge-based image restoration algorithm. For restoring image structures, we first use the segmentation result with iterative Otsus thresholding to obtain the information of edges. Bezier curves are then used to reconstruct the image skeletons in missing areas. It improves the limitation of the edge-based restoration approach which approximates incoming edges with only lines and circle arcs. The inpainting process is divided into two phases: the first phase restores the image structure by pixel-based interpolation, while the second phase fills holes for preserving texture information with patch-based inpainting method. Experimental results on both synthetic and real images demonstrate the effectiveness of the proposed method. By restoring the curvature structures and textures of the damaged regions, the proposed method achieves content-aware image inpainting.

Keywords: image inpainting, texture synthesis, exemplar-based method, edge structure reconstruction, pixel-based interpolation, Bezier curves

Full Text () Retrieve PDF document (201207_08.pdf)

Received July 14, 2010; revised October 15, 2010; accepted December 30, 2010.
Communicated by Chung-Ling Huang.
* This work was supported by the National Science Council of Taiwan under Grant No. NSC 97-2221-E-005-081.