"Use Sprite Mesh" option recently added to UI Image - What is this for exactly?

The documentation for it is here:
https://docs.unity3d.com/ScriptReference/UI.Image-useSpriteMesh.html

When I first read it I assumed it's use case was for non-square clickable UI to make the clickable area of the button match the shape instead of always being a square, which would have been a really cool new feature, but after testing it doesn't seem like this is the case.

So, if this option isn't for making the clickbox of a UI image fit the shape, what is it for exactly?
Just to reduce overdraw?
If it's usage case is performance based, should I be using this on UI images with lots of whitespace?

Or am I just doing something wrong? Is anyone else able to get this option to work for making irregular shaped clickboxes for irregular shaped UI buttons?

I hate to have to ask this here, but it seems like there is almost zero information on this feature, (which I believe just showed up in 2018.3) and it seems like it could be a useful feature if it's intended usage was known.

 

I would also like to know more information about this.

 

The diference between using this feature or not is as showing in the attached image.
At left, Use sprite mesh is off (generates a mesh with only two triangles, as always behaviour).
At right, Use sprite mesh is on (generates more triangles, same as using a sprite renderer).

So, the generated mesh will react to clicks. Notice that the sprite mesh generated (at right) does not fit the sprite alpha.
To use the Use sprite mesh feature, you should configure at the sprite import settings > Mesh type > Tight

This feature is usefull for UI optimization purposes, specially for large images such as backgrounds.

 

If you're looking for a way to do hit-detection on the shape of an image, for instance a UI Button, you can use this class: https://github.com/lucbloom/unity_image_with_hit_detection_on_shape
It takes the triangle mesh from the sprite texture and matches it with the exact coordinate of the mouse input on top of it.
Instructions in the readme.md

 

Hi,
How does it help for UI optimization while It's mesh has more vertices.

 

AFAIK, with 2 triangles we are using transparency which is bad, if we are using more tris but with no transparency, it should be cheaper