Depth guided image completion for structure and texture synthesis

Michael Ciotta; Dimitrios Androutsos

ICASSP 2016accepted0 citations

Depth guided image completion for structure and texture synthesis

Abstract

This paper presents a method that separates the image completion process into structure and texture synthesis. A method is first introduced for completing the respective depth map through the use of a morphological diffusion-based operation, ensuring that structure is propagated smoothly across the unwanted region. These estimated depth values are then used to guide completion of the missing color image texture. A fast exemplar based PatchMatch approach is adopted and an extension of the coherence-based objective function introduced by Wexler et al. is used to complete the resulting texture in both the color and depth images. Finally, we provide experimental results to demonstrate the superiority of our method against others in a variety of different scenes.

BibTeX

@inproceedings{icassp2016_depthguidedimage,
  title = {Depth guided image completion for structure and texture synthesis},
  author = {Michael Ciotta and Dimitrios Androutsos},
  booktitle = {ICASSP 2016},
  year = {2016}
}