Developer Reference for Intel® Integrated Performance Primitives 2019
Adds the offset value after filtering operation for ipp8u and ipp16u data types , and sets the rounding mode.
IppStatus ippiFilterBorderSetMode(IppHintAlgorithm hint, int offset, IppiFilterBorderSpec* pSpec);
ippi.h
Headers: ippcore.h, ippvm.h, ipps.h
Libraries: ippcore.lib, ippvm.lib, ipps.lib
hint |
Suggests using specific code for rounding. Supported values:
|
||||
offset |
Constant that is added to the final signed result before converting it to unsigned for ipp8u and ipp16u data types. |
||||
pSpec |
Pointer to the initialized filter specification structure. |
This function adds the offset value after filtering operation for ipp8u and ipp16u data types with the ippiFilterBorder function:
pDst=(summ(src[i]*kern[i]))+offset
You can also use this function to set the rounding mode for the filtering result.
The 8u_C1R, 8u_C3R, 8u_C4R, and 16s_C1RFilterBorder function flavors initialized with the ipp16s coefficients support ippAlgHintNone and ippAlgHintAccurate rounding modes.
Use this function after the ippiFilterBorderInit function and before calling ippiFilterBorder.
ippStsNoErr |
Indicates no error. |
ippStsNullPtrErr |
Indicates an error when one of the specified pointers is NULL. |
ippStsNotSupportedModeErr |
The offset value is not supported (for ipp16s and ipp32f data types). |
ippStsAccurateModeNotSupported |
The accurate mode is not supported for some data types. The result of rounding may be not exact. |