Super-resolution (SR) are techniques that in some way enhance the resolution of an imaging system
Michael Smith at Carlenton University has created a
C++ implementation of the superresolution algorithm
(his program executes demosaicing too).
Visilab Reseach Center has been able to run the
Smith's algorithm on a embedded processor (Sony
PSP), thank to the features of Nanodesktop 0.3.8.
As you can see, a detail of an image (size 135x111
pixels) has been enlarged in the output image (2x
superresolution factor).
But, in the 270x222 pixel image, there are MORE
details (see the faces of the children: in the input
images, you see the pixel effect; in output image
you cannot).
It is advisable to see this video in HQ mode.
Link to this comment:
All Comments (0)