Here is a nice paper from the last SIGGRAPH 2017 about Post-Capture Image Composition:
http://cvc.ucsb.edu/graphics/Papers/SIGGRAPH2017_ComputationalZoom/
Here is a nice paper from the last SIGGRAPH 2017 about Post-Capture Image Composition:
http://cvc.ucsb.edu/graphics/Papers/SIGGRAPH2017_ComputationalZoom/
I use this for construction in 3D http://www.gcc.tu-darmstadt.de/home/proj/mve/
makescene -i . scene
sfmrecon scene
dmrecon -s2 scene
scene2pset -F2 scene temp.ply
fssrecon temp.ply temp1.ply
meshclean -t10 temp1.ply output.ply
Afterwards can use meshlab or blender to create new images