Add docs (tutorial) about transfer image from RAM to pipeline (model)

**Is your feature request related to a problem? Please describe.**
When I started learning ML.NET, it was hard to understand how to adequately transfer an image from RAM (byte[], SKImage, MLImage) to a pipeline (model) (which automatically resizes the image and classifies it). I couldn't find this in the documentation, and I couldn't find any clear ways on the Internet.

**Describe the solution you'd like**
Maybe you could add information about this to the documentation?
I have some code that just needs to be adapted for use in documentation: https://github.com/Lavshyak/MLNetImageFromRamStrange/tree/master/ImageFromRamStrange
And it's worth mentioning, it was a non-intuitive moment: https://github.com/Lavshyak/MLNetImageFromRamStrange/blob/788d67d05e0ebb00d4db0753ac3298a9507a5805/ImageFromRamStrange/DataModels.cs#L12
And may be you can copy some code from there: https://github.com/Lavshyak/MLNetImageDisposingProblem/blob/master/MLNetImageDisposingProblem/Reproduction.cs

And don't forget about that): https://github.com/dotnet/machinelearning/issues/7444