North is approx. towards the bottom of the image.

The three biggest visible galaxies are part of the same galaxy group but are too far apart from each other to be interacting. From left to right, we see:

NGC 5981 is a spiral galaxy of Hubble type Sc at a distance of about 86 million lightyears, which makes its size about 65000 lightyears. We see it almost edge-on and the central dust band is nicely revealed. I found few studies about its global properties, although its star formation regions have been used in many as an example of an undisturbed/typical star population.

NGC 5982 is an eliptical galaxy and the brightest of the group whose namesake it is. Its nucleus is listed as a LINER-candidate. Its distance is approximately 132 million light years, which makes its size about 100000 light years. It is a "shell-type" galaxy, which means its types of its stars are distributed in distinctive shells around its core. It is thought that this is caused by periodic gravity waves after a galactic collision.

NGC 5985 is a grand design barred spiral of Hubble type SBb. Its distance is about 120 million light years which makes it a huge spiral almost twice the Milky Way in size. Other than size or distance, information on it is sparse.

On the left hand side there is also the small spiral NGC 5976 which is part of the same galaxy group. Numerous other background galaxies can be seen in this image.


I've processed this image in two different ways, because I am very sceptical about the use of AI-based sharpening and noise reduction techniques for astronomical images. The "final" version ist processed with my usual workflow, while the AI version has Topaz Denoise and Topaz sharpening applied to the luminance channel. The color channels are the same between the revisions, although in both cases I used a final LocalHistoryNormalization on the combined image. My usual workflow is: Noise reduction with Mure (scale 0.5), TGV and MLT to the linear data and slight ACDNR to the stretched one, all applied to the luminance. Sharpening/contrast enhancement by Deconvolution, MLT and LocalHistoryNormalization. For the AI revision I've used Mure Denoise (scale 0.5) and immediately stretched, then imported into the Topaz filters. I've then blended them with the original data at about 2:1. 

Personally I do not think the AI-version has added detail that's not present in the data. Instead, it has reduced the size of the star forming regions in NGC 5985 to be more like strands instead of cloud-like features. I could not tweak the parameters to get rid of this effect. The difference is, admittedly, subtle, but keeps me from going all-in with the AI-based filters. The time I needed to sharpen/denoise was very notably different:  I spent about 10 minutes on the AI image and 2 hours on the non-AI version, though. The lure of the quick way is always strong for the Padawan...