Luminar Neo is an advanced photo editing package from Skylum, the creators of the Luminar range of software including Luminar Neo which was released around Feb 2022.
Whilst the software was released around Feb 2022, it was not feature complete, and was missing a number of features planned to be added in a future update. One of these features was Portrait Background Removal AI which this post explores.
Portrait Background Removal AI is a tool that uses AI to identify people in the photo and select those people to retain them, whilst masking/selecting and removing the background. This aims to speed up the process by automatically doing most of the masking, whilst allowing for some minor selecting and tweaking of the AI selection/mask.
As I am not much of a portrait photographer, I actually didn't have many examples I could use to test this software (since I focus more on wildlife, macro, nature and environmental photos) so I found some pictures on PixaBay which I have included below.
In this test, I used default settings and did not do any additional tweaking of the mask so I could show by default what the software masked in and out, it should be noted though that the software does allow for tweaking the mask so I am sure with some time and tweaking I would be able to get these photos masked perfectly.
I have used image compare modules, so the image on the left is the original image, whilst the image on the right is the image with the background removed using the Portrait Background Removal AI tool and just the default settings.
Example Photos
Image Sample 1: Image by Shahid Shafiq from Pixabay
Image Sample 2: Image by adamkontor from Pixabay
Image Sample 3: Image by Carlos Alvarenga from Pixabay
I hope this gives you a little idea of how the software works under default conditions with no tweaking, clearly, I could have done some tweaking the mask to get a better result but I wanted to show how it would go given a different set of images. I personally think it did a pretty good job, and the images would just need some tweaking to fix the mask. It stands to reason that the more complicated the image, the harder the mask is to get right, which is what appears to be happening around hair and other more difficult features. It should be noted that this test was done using a pre-release piece of software, so it is possible that performance may be improved in the final release.
If you are interested in finding out more about Luminar Neo, you can click on any of the links in the post, or you can find out more by clicking here. If you are interested in purchasing make sure to use the coupon code travishaleneo at checkout and you may get a discount on your purchase.
Comments are closed.