Extending ImagePoints for more informative Visualization

I’ve submitted an initial PR to get a thumbs up/thumbs down on adding a class to improve ImagePoints visualizations.

I have documentation written, and tests written locally, and have still have to blend the formatting, but this first commit is to get feedback on the code.

Here’s a “doc_src”-type notebook demonstrating the uses of the new class, MultiIp in vision.image module, and further discussion about dev issues.

Also, open to suggestions for renaming the class, maybe MultiImagePointsDisplay?