Template Matching using SqrDistanceFull_Norm

Where is the documentation on this and/or a sample that tells me how to use this function. For instance pDst Destination-Image Pointer. I guess this is an image pointer that contains the results from the template match. What are it’s dimensions, how should I create it, what type does it contain and what do the results mean? All I see if a function prototype that I can guess what the parameters are for?

SqrDistanceFull_Norm

The functions compute the \sigma_{st}(c,r) in General Introduction with full mode (see Categorizations).
Common parameters for nppiSqrDistanceFull functions include:

Parameters
pSrc Source-Image Pointer.
nSrcStep Source-Image Line Step.
oSrcRoiSize Region-of-Interest (ROI).
pTpl Pointer to the template image.
nTplStep Number of bytes between successive rows in the template image.
oTplRoiSize Region-of-Interest (ROI).
pDst Destination-Image Pointer.
nDstStep Destination-Image Line Step.
nScaleFactor Integer Result Scaling.
nppStreamCtx Application Managed Stream Context.

Most npp functions are not documented with anything remotely resembling a full example code or detailed descriptions. The few exceptions would be those that are demonstrated via CUDA sample codes, and some worked examples posted on this forum.

The npp library mimics behavior in the ipp library. You can occasionally find ipp sample codes that are instructive.

image distance function behavior is generally described here:

https://docs.nvidia.com/cuda/npp/group__image__proximity.html#general_introduction

image storage, and ROI handling is generally covered in various places in the npp docs, and there are various CUDA sample codes that demonstrate how to pass an image to a npp function and how to specify source, destination, ROI, etc.

You’ll need to do some research and study. If you’d like to see a change to CUDA documentation, you are encouraged to file a bug requesting that. The instructions for bug filing are listed in a sticky post at the top of the CUDA programming forum.