Abstract
Plenty of complex image editing techniques require certain per-pixel property or magnitude to be known, e.g., simulating depth of field effects requires a depth map. This work presents an efficient interaction paradigm that approximates any per-pixel magnitude from a few user strokes by propagating the sparse user input to each pixel of the image. The propagation scheme is based on a linear least squares system of equations which represents local and neighbouring restrictions over superpixels. After each user input, the system responds immediately, propagating the values and applying the corresponding filter. Our interaction paradigm is generic, enabling image editing applications to run at interactive rates by changing just the image processing algorithm, but keeping our proposed propagation scheme. We illustrate this through three interactive applications: depth of field simulation, dehazing and tone mapping.
Files
Source code
The source code for this paper has been released to the public domain.
Video
BibTeX
Acknowledgements
This research has been partially funded by the European Research Council (ERC) under the EU Horizon 2020 program (CHAMELEON project, grant agreement No 682080), the Spanish Government (projects DPI2015-65962-R, DPI2015-69376-R) and Aragon regional government (Grupo DGA T04-FSE). The authors would like to thank D. Gutierrez and J.J. Guerrero for their support on this project.