Okay, let's see if I got your question right...

A point cloud will not have a "resolution" the way a raster does simply because all the points will not be a an equal distance from each other. For instance, there can be an area of your model where you have, say, 100 points/square cm, and another where not as many photos were taken where you'll only have 10 points/square cm. Your point cloud density will also *greatly* vary depending on the parameters you chose for the alignment (no. of photos, tie points, key points, etc) and the dense cloud generation (high density, medium density, etc).

If you want to get an average density of points, you could take the number of points you have an divide it by the surface area of your model.

As for accuracy, that's an *entirely* different can of worms.

Cheers