I am looking at the general pyramid sample in the documentation, and was able to get the code running - so I have both the gaussian and laplacian pyramid of the input img. Now I want to rebuild the original image using the gaussian and laplacian pyramid using the formula: Original image = G0 = L0 + pyrup(G1).
// both gaussian and laplacian pyramids have been constructed. Only diff from sample is this is using a 3-channel img int tempStep; Ipp32f* temp = ippiMalloc_32f_C3(width, height, &tempStep); status = ippiPyramidLayerUp_32f_C3R(gImage, gStep, gPyr->pRoi, temp, tempStep, roi, *gState); int tempStep2; Ipp32f* temp2 = ippiMalloc_32f_C3(width, height, &tempStep2); status = ippiAdd_32f_C3R(temp, tempStep, lImage, lStep, temp2, tempStep2, roi);
For the ippiPyramidLayerUp function, I wasn't sure if I should use gState or lState for the final parameter. I have tried both and they both give me the same result and the final output matches the input image - so can I assume it doesn't matter if I use gState or lState?