Implement orthogonal resampling as a convolution

[kabasset]

Interpolators could be kernels, too, when the window will be part of the filter operator.
Then, separable interpolation could be much accelerated.

[kabasset]

Any orthogonal affinity could be applied this way, i.e. if limited to translation and scaling. To be dispatched directly in warp().

[kabasset]

It is also possible to make rotations semi-separable, using a sequence of shears, but not sure we would save time.
As for the integral-factor resampling, this is a true convolution.