I use the Skylum range of photo editing tools on a daily basis (from Aurora HDR to Luminar AI), and as an affiliate of their program I received early access to a technology preview build of “Luminar Neo” to test out the DustAI and PowerLineAI Removal tools. For those who are not aware Skylum are the creators of a range of software utilities / editors including Aurora HDR, as well as Luminar AI and some time ago they announced (to some criticism) their upcoming Luminar Neo program. Whilst a full description of all the packages is outside of the scope of this article, Luminar AI is a photo editor which is designed to help people make simple edits to their photos, and focus on the results (if you were to compare it would be akin to Lightroom). On the other hand, Luminar Neo is more about making more complex edits, using layering, and rebuilds the underlying engine completely to support more complex edits (where Luminar Neo is more akin to Photoshop). Whilst I have announced a number of different features, I have until now only been provided with examples from Skylum to demonstrate some of the features. This however, recently changed when I was provided access to a technology preview which allowed me to use my own images to demonstrate some of the limited features which are demonstrated in the technology preview.
This preview demonstrates a number of features, but in this article I am looking at Dust Spot Removal AI and Powerline Removal AI. These tools are fairly self-explanatory where dust spot removal uses AI & machine learning to identify dust spots in the image and remove them, whilst Power Line Removal does the same but removes power lines from the image.
In the demonstration I used an image of the sky (below), in the first image I have put red circles around the dust spots to highlight the dust spots (and their intensity).
In the second image (below), I have natively applied the Dust Spot removal (with no other editing) and it has done a great job of removing the dust spots, my only criticism is that the spot slightly left of centre (which sits in the middle of the powerlines) has been removed, but it has not stitched the power line back properly so you can see the power line is a bit disjointed (although to be fair this would be easily fixed with some manually cloning).
In the third image (below), I have applied the Dust Filter AI and Powerline Removal AI with just the default settings which let to the image below. As you can see it has removed lots of the power lines but in reality it has struggled with a few of the power lines in the centre left, and centre right of the screen. In reality this was a fairly easy fix and I could clone these out leading to the fourth image.
The fourth image (below) was with the Dust Removal AI and the Power Line Removal AI applied, but then I also manually cloned the areas where the power lines were not completely removed to fix it up. This only took about 5 seconds of additional work just to get rid of the power lines which were remaining.
Whilst I have only used this on a few images, I did note that it seemed to work fairly well. There can be some minor issues in complex images where the dust spot has been removed but there is a stitch error where it has been cloned (i.e. around the power lines). This was not a major issue and just needed some manual correction but I think it shows that the underlying technology seems to work fairly well, and is far easier to do than manually looking through the image and finding each of the spots and removing them one by one. In addition the power line removal does a fairly decent job but again it is not foolproof so you do need to double check and do any manual removal as well. One thing I did note with respect to the powerline removal is that it only removes the power lines, if there are any power poles in the image you will need to manually remove those by yourself.
I intend to do some videos in the near future, so stay tuned to my website and YouTube channel for updates. In the meanwhile if you are interested in finding out more about the Skylum Luminar Neo package, you can do so by clicking on any of the links in this post or by clicking here.
Comments are closed.
PO Box 113, Beaconsfield Upper, Victoria, Australia, 3809.