| | |
Summary: A Framework for Content-Adaptive Photo Manipulation Macros:
Application to Face, Landscape, and Global Manipulations
Supplemental Materials
page 2-3 Tilt Shift Manipulation Results
page 4-5 Car Recoloring Manipulation
page 6 Brushing without Landmark Points
page 7-10 Mechanical Turk Results
page 11 Parameter Value Comparison
page 12 MSE Results
page 13 Least Squares Comparison to LARS
Tilt-Shift Manipulation
We demonstrate the tilt-shift manipulation on 10 images, and our framework successfully learns to blur the
regions above and below the car.
Target Images Macro Results
...
Demonstration Manipulations:
Target Images Macro Results
Car Recoloring Manipulation
We demonstrate recoloring of 5 red cars and 5 green cars to blue. Our content-adaptive macro correctly recolors new
target cars, without recoloring red/green elements (e.g. red owers) that fall outside the car bounding box. But, it fails
|
