calculate non-local median of image windows
Use image windows of given size, calculate non-local median within this window and use the result as new value for the window center pixel. This eliminates outliers and makes e.g. flow estimation more robust. Median calculation is being performed using the "Iteratively reweighted least squares" approach (e.g. "Numerical Methods in Scientific Computing Vol. II" by Germund Dahlquist and Åke Björck, SIAM, section 8.7.5 )The non-local approach was introduced by D. Sun within his 2010 CVPR paper "Secrets of optical flow".
input slots | |
InputSlot< CImgList< T > > | img |
image data input slot | |
InputSlot< CImgList< T > > | motionUV |
weight input slot | |
output slots | |
OutputSlot< CImgList< T > > | out |
image data output slot | |
parameters | |
Parameter< unsigned int > | iterations |
iteration count | |
Parameter< T > | sigma_color |
color difference weight | |
Parameter< T > | sigma_occ_color |
occlusion color diff weight | |
Parameter< T > | sigma_occ_divergence |
occlusion divergence weight | |
Parameter< T > | sigma_spatial |
spatial difference weight | |
Parameter< unsigned int > | windowRadius |
radius of image windows (size is ) |
This Module subclasses TemplatedParameteredObject< T >.
For documentation of parameters and slots inherited by this base class,
please have a look at the corresponding Module reference.
This module is templated. There are instances with T=int,float,double.
An additional parameter called templatetype
may be used to select which instantiation you want.
The detailed doxygen documentation beyond the parameters/slots may be found here.