GMM Adjustment

GMM Adjustment

Hello friends. Could you provide correct parameters for Gaussian Mixture  (ippiForegroundGaussianInitAlloc parameters).  I'm going to use it in indoor place with constant illumination. I have already tried to use GMM with extensive diapason of parameters, but it works unsatisfactorily. For example: Shadows aren't detected, therefore foreground is fail.

Thanks.

3 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.

Do you mean input arguments for the ippiForegroundGaussianInitAlloc function? If Yes, then:
...
// Arguments:
// pSrc Source image
// srcStep Step in source image
// roiSize Source image ROI size.
// pModel Pointer to the structure of the gaussian model.
// pState Pointer to the pointer to the segmentation state structure.
...
IPPAPI( IppStatus, ippiForegroundGaussianInitAlloc_8u_C1R, ( const Ipp8u* pSrc, int srcStep,
IppiSize roiSize, IppFGGaussianModel* pModel, IppFGGaussianState_8u_C1R** pState) )

IPPAPI( IppStatus, ippiForegroundGaussianInitAlloc_8u_C3R, ( const Ipp8u* pSrc, int srcStep,
IppiSize roiSize, IppFGGaussianModel* pModel, IppFGGaussianState_8u_C3R** pState) )
...

Or, Do you mean parameters to initialize a Gaussian Model ( IppFGGaussianModel )?

>>... I have already tried to use GMM with extensive diapason of parameters, but it works unsatisfactorily. For example:
>>Shadows aren't detected, therefore foreground is fail.

Please provide more technical details. Thanks in advance.

Hello Sergey. Yes, I mean values of variables in IppFGGaussianModel. Such as priorBack and others, excluding numGauss.  I suppose, that I set  wrong values. I need a template for comparison. Can you show me good values (good profile) for correct work of GMM?

Great Thanks.

 

Leave a Comment

Please sign in to add a comment. Not a member? Join today