OK, this is going to really get my goat.
Except that this test image has just a face and part of a shoulder, without any naughty bits. Not even erotic at all.
Except that everybody KNOWS that it's a cutout of some porn. You know what? That's tasteless and tacky. Years ago when I was a PhD student in computer vision (more later) I used to help supervise an image processing practical. It used Lena of course. Sooner or later someone found out and everyone had a good laugh.
Well the guys in the practical (90% of the people). The other 10% didn't seem to think i's so funny. I can't imagine why. So yes having a PlayBoy centrefold, even cropped sets completely the wrong tone. It sets the tone that it's no just male dominated but a total boy's club too.
For some reason all the other images are bland and neutral. You never see someone pointing out how good the reproduction of high frequencies on Baywach era David Hasslehof's manly chest or anything.
It's always Lena, peppers, and a few others.
It's a good test image because it catches both distortions of detail and color damage to areas with a gentle gradient.
No, it was good perhaps 15 years ago. It has a little high frequencies and a bit fo gradient. There are far beer trickier images now. There are far beer images it's easy to get hold of that are much harsher in those regards. All the massive overuse of Lena does is lead to ad-hoc reverse engineering of tha picture. Or as some wag in the field pointed out, the Lena compression algorithm: If the first bit is 0 then emit Lena, otherwise decode any remaining data as a JPEG.
If I have to see yet another buchered version of a Lena picture again in a compuer vision paper that I read or review, I seriously think I will murderize someone.