Are Matlab parameter sets even interpreted?

[Harzknopf]

differing parameter sets for mex_SparseRLOF(...) produce exactly the same flow fields. Is that even possible?

[Tobias Senst]

Maybe but can you send me the parameters so that I am able to check them

[Harzknopf]

These are the parameter sets:

parameter1 = struct('maxLevel', 5, 'maxIter', 30, 'HampelNormS0', 3.2, 'HampelNormS1', 7, ...
'LargeWinSize', 21 , 'SmallWinSize', 9, 'MinEigenvalue', 0.001, 'RansacReprojThresholdPercentil' , 71, ...
'SegmentationThreshold', 30, 'useIlluminationModel' , 1, 'useGlobalMotionPrior', 0, ...
'SupportRegionType', 'SR_FIXED', 'SolverType', 'ST_STANDART');

parameter2 = struct('maxLevel', 5, 'maxIter', 20, 'HampelNormS0', 3.2, 'HampelNormS1', 7, ...
'LargeWinSize', 21 , 'SmallWinSize', 9, 'MinEigenvalue', 0.001, 'RansacReprojThresholdPercentil' , 71, ...
'SegmentationThreshold', 30, 'useIlluminationModel' , 0, 'useGlobalMotionPrior', 0, ...
'SupportRegionType', 'SR_FIXED', 'SolverType', 'ST_STANDART');

[Tobias Senst]

Thank You! This is a bug in the matlab wrapper parameter interpretation function.
If you can try to bipass this by using the python or c++ version interfacs. I will look at this issues, but it could take some time.