Image
Removing the image (and eventually replacing it hopefully) is fine by me. That part's simple.
Types
In terms of splitting into types, I'm rather neutral. I do feel this effectively serves the purpose of what's already listed on the page, where Invulnerability against all conventional attacks with some exceptions is the standard, and Invulnerability against specific ones is Limited. My concern is more if this would be worth the massive amount of work to re-categorize since it doesn't fundamentally change anything but does require specification on all existing profiles with the ability.
Name
What I do have strong thoughts on is the name. I think it's greatly underestimated how much the name of an ability impacts one's perception of it, even if you do take the time to try and read and parse the entire page. Furthermore, fiction uses this term all the time in a completely different meaning which is downright confusing for anyone not super familiar with our standards.
We serve a wide user base, and while certain abilities do require explanation for full understanding, it's my opinion that it's a no-brainer for accessibility and to prevent confusion (which I see a LOT in threads, Invulnerability is one of the top abilities I have to reject and explain to people) to make the name as close to the definition per common use as possible.
From that perspective, renaming it to something like Damage Nullification makes it far more specific to us (so there's less confusion over the way most media actually uses the term), and a lot more self-explanatory for what we actually mean, in terms of immunity as opposed to just reduction or high durability. I genuinely think this would make it easier to understand and more likely to be applied correctly.
Still, 'Invulnerability' as a term isn't completely nonsensical either, which to me means we don't necessarily have to go through and scrub every use of the term so long as we can just have it link to Damage Negation and use that terminology going forward, and changing it when convenient.