Object rearrangement consists of moving an object to a new image location, or deleting part of the image, while keeping some of the content of the image unchanged. The user selects a region to move, and specifies the location at which the selected region will be placed. A new image is generated satisfying this constraint.
Object rearrangement is specified in two constraints in the shift-map formulation. One forces pixels to appear in a new location (marked in red or in green when the location doesn't change, hover with the mouse over the Shift-Map result to see the constraints on the output) and the second marks these pixels for removal from their original location to avoid their duplication in the output
Here is a comparison of our method with the Patch-Transofrm method (see reference below):
|
Input Image |
Patch-Transform Result |
Shift-Map Result |
|
|
|
|
|
|
|
|
|
|
|
|
| In this example we have also added a constraint to remove one of the objects in the input: |
||
|
|
|
|
