Mat vs Image<,> pixel access Get Set

[abibibu]

Hello, I'm trying to figure out the best practice for using emgu.CV, I see in this tutorial that the newer versions of openCV is based on mat and from some test that I have run I see that Mat performance seems better, in my application I need to access single pixel very often (i.e. rebinning, interpolation, convolution, remapping) with images that operation is simple using Image<,>.Data[,,] but with mat I dont know how to do that without reallocating the memory.

I see from this that openCV has a specific template for pixel access.

Should I use Images and forget about mat or does a pixel access method exists?

thanks.