Choosing the best photos from those we took is one of the most pleasant and at the same time the most tedious activities that a photographer has to deal with. It's time to make her comfortable.
Rating photos by AI algorithms is something you've already heard of for sure. Attempts have been made to select photos for competitions, for our own satisfaction so that we can see how good we are photographers. The usefulness of the effects of such selection is a wide topic, because despite the growing advancement of AI algorithms, still assessing the aesthetic value of photos seems to be beyond the reach of objective assessment by the machine. Technical correctness can easily be inferred from photo content analysis, image quality estimates, but that's not all that makes photography attractive. The problem (still) of AI algorithms is precisely their objectivity and that is why we still use the image viewing application ourselves.
And when evaluating photos, a dose of subjective approach to the topic is also useful. Because AI is still difficult to understand that the world of photography is not a book. Every artist, and often photographers are considered to be such, has his own style, a specific view of reality, which can contradict generally accepted rules, and at the same time be a perspective that gives the photos freshness and diversity.
When we extend the photographer's classification to everyone who takes pictures, we attach a smartphone photography (without taking anything of its value), before the AI the task will be even more complicated. And yet software developers are trying to entwine in applications that support AI spirit cameras, which will help in creating the correct shots and eliminate the need for subsequent selection.
Such solutions can be noticed especially in the mobile industry in the form of the function of automating the selection of frame by built-in AI smartphones. AI digital cameras also appear more and more often, but more for the purpose of correctly detecting frame elements than suggesting the right shots.
Sorting photos in Adobe Lightroom Classic
Computer photography applications are different. Here we have a wider field to show off. Canon wants to use it, which for the popular photo manager application and digital darkroom Adobe Lightroom Classic prepares a special Photo Culling plugin. Its task, as the name indicates, is to verify photos and determine their usefulness.
Photos will be divided into better and worse. This will save time, which will be especially appreciated by professional photographers, who often have to quickly choose from the large collection of photos the most interesting (remember this last word). An ambitious task, but if the Photo Culling plugin succeeds, then not only professionals but everyone will benefit. In an era when time is a very valuable value, such use of AI will positively influence its perception as a useful technology.
Criteria for the selection of the best photos by AI Canon
The key word in this case is the word "most interesting". For the algorithms developed by Canon, this will mean a selection based on patterns that take into account sharpness, noise level, correct exposure and arrangement of motifs in the frame. Also, small things like open or closed eyes are to be a criterion for positive or negative selection.
We hope to see the effects of this plug-in soon, as it is expected to hit the Adobe Exchange App Marketplace repository in the first months of this year. Then we will learn how important the word Canon is when determining the creator of the plugin, as well as how much AI we use in it to conduct. This is important, because if it turns out that the program does not leave us any freedom in determining the parameters of the "most interesting" photos, then the whole undertaking in my opinion will be just art for the sake of art.
Source: Canon, photo: Karol Żebruń
More about applications of Artificial Intelligence:
I’m kris. founder of pop-software.net
I am a freelancer, programmer and expert on windows / linux systems. I work at a computer company in India.