Create an image manipulation tool that allows assigning 3 dimensional geometry to objects in the 2D image, manipulating those objects in 3D, then re-rendering the image.
For small manipulations, the assignment of geometry does not need to be very accurate. Computer vision tools could give suggestions. Not only geometry but also BRDF of materials and locations of light sources. Moving an object in 3D may cause it to go more in or out of focus.
Designing the UI seems very challenging.
No comments :
Post a Comment