This is a modified mandelbrot. Maybe distorted by the used software. Look at the minibrots. If this would be the standard Mandelbrot, they would be symmetrical and identical to the large one.

This is my reasoning so far:

The outer light filaments have 3-way spirals.  So the first angle is {1,2}/3.
The inner light filaments have 18-way spirals.  So the next angle is {1,5,7,11,13,17}/18.
There are dark embedded Julia sets visible in the light filaments, so the zoom must have passed an island near its seahorse valley (because of the shape of the EJSs).
There is no additional shape stacking, so the zoom passed at most one island.
The thick filaments are tuned by an island.
The outer thick filaments have 2-way spirals.  So the first angle from the island is 1/2.
The inner thick filaments have 11-way spirals.  So the second angle from the island is {1..10}/11.
The largest visible island is highly distorted, which occurs at the lowest period island in the widest spoke attached to the principal hub.
So the angled internal address of this distorted island detuned relative to its parent is 1 1/2 2 a/11 22 1/2 24.
And the angled internal address of the parent island begins 1 b/3 3 c/18 54 1/2 ....

Attached is not quite right, probably incorrect choice of numerators and the missing ... in the parent address

Location for MuET for FragM:
Zoom = 123456789 Logarithmic
Center = -0.05794121488070425808200419,0.8132682994364264091896378

Wow !!!!

 Thx

related ( but not so complicated ) :
https://math.stackexchange.com/questions/1093525/finding-the-location-of-an-image-of-the-mandelbrot-set

Zoom = 1/ radius
Needs double precision, can't remember if my published muet.frag supports that.

Should work in other Mandelbrot viewers too, but muet has the right gamma correct colouring for this effect. I set Exposure to 2.5.

Increasing iterations dos not improve image

image is fine, just colouring is not so nice - m-perturbator-gtk is optimized for annotations so contrast is low. I should add an exposure control to adjust filament thickness but that needs supersampling to look good.

m-perturbator-gtk uses more precision for the reference to make zooming in smoother (it can reuse calculations)

The precision needed for plain rendering without perturbation depends on pixel spacing which can be calculated from radius and image dimensions. Radius of this image should be fine with double for screen sized images.