Use mri-stack from TestImages. Fixes #40

[timholy]

Requires https://github.com/JuliaImages/TestImages.jl/pull/29 and https://github.com/JuliaGL/GLVisualize.jl/pull/143. The new file is quite small, so the resolution isn't impressive, but on the other hand the anisotropic pixel spacing represents a good test of functionality that's crucial for almost anyone doing volumetric visualization.

[timholy]

This can be closed, right?

[SimonDanisch]

Or merged?

[timholy]

I don't think there's anything important here anymore...the main problem this was designed to solve was the fact that the original example referred to a file on your harddrive, and I was trying to use a publicly-available dataset. You replaced this file with one where you synthesize a 3d dataset, and in some ways that's nicer because you have complete control over resolution, etc. So perhaps the only reason to use this is if you want users looking at a "real image" rather than a synthetic one.

One other point worth making is that "mri-stack" has very poor resolution along the z axis, and it's 5x different from the resolution in x and y. That might be useful in terms of making sure that your display algorithms handle anisotropic images correctly; on the other hand, it looks really crappy if the user rotates the sample so that the z-axis is nearly in-plane.

[timholy]

Let me know what you want to do; if you decide you want this (perhaps as a second volume example), I'd be very happy to contribute it.