Object-based image inpainting

作者: Feng Wu , Xiaoyan Sun , Shipeng Li , Dong Liu

DOI:

关键词: PixelMatching (graph theory)InterpolationVisual ObjectsArtificial intelligenceSegmentationObject (computer science)MathematicsDeep-sky objectPattern recognitionInpaintingComputer vision

摘要: Systems and methods provide object-based image inpainting. The inpainting can recover part of a missing object or synthesize visual objects that are entirely missing. In one implementation, an exemplary system generates target to be inpainted into image, by geometrically matching key features with model selected from the same another image. Borders refined through pixel-based contour deformation textures for obtained either pixel interpolation patch-based modeling pre-existing texture in Graph-cut segmentation rectifies overlap as is embedded Poisson editing filters seams, edge filling completes gaps around border object.

参考文章(11)
Michel Gangnet, Patrick Perez, Kentaro Toyama, Andrew Blake, Antionio Criminisi, Image region filling by exemplar-based inpainting ,(2004)
Mark Hamburg, Todor G. Georgiev, Jen-Chan Chien, Generating replacement data values for an image region ,(2002)
Jhing-Fa Wang, Han-Jen Hsu, Shang-Chia Liao, Image-capturing device and method for removing strangers from an image ,(2005)
Tony F. Chan, Jianhong Shen, Nontexture Inpainting by Curvature-Driven Diffusions Journal of Visual Communication and Image Representation. ,vol. 12, pp. 436- 449 ,(2001) , 10.1006/JVCI.2001.0487
Lingxiang Zhou, Donghui Wu, Yiqing Jin, Yunfang Zhu, Huiqiong Wang, Texture synthesis for repairing damaged images ,(2003)
Alexei Kotlikov, Krzysztof Zaklika, Method for removing defects from images ,(2001)