Jordan News | Latest News from Jordan, MENA
August 13 2022 2:42 AM ˚
e-paper Subscribe Sign in My Account Sign out

AI helps customize computer graphics

DALL-E-2 ai
(Photo: Twitter)
  • +
  • -
DALL-E-2 ai

Jean-Claude Elias

The writer is a computer engineer and a classically trained pianist and guitarist. He has been regularly writing IT articles, reviewing music albums, and covering concerts for more than 30 years.

One of the most striking contributions of Artificial Intelligence to cutting-edge computer programming is architectural visualization. It can be seen and experienced in, among others, two new applications that are nothing short of mind-blowing, even if you you think you have seen it all.اضافة اعلان

The first is called Canvas and is the brainchild of Nvidia, the world’s leading designer and maker of graphic processors for computers. It is difficult, in words only, to give a complete idea of what Canvas can do. A better way would be to watch the videos available on YouTube on the subject.

Even Nvidia’s own definition comes short of describing the application’s spectacular work; it “lets you customize your image so that it’s exactly what you need. Canvas has nine styles that modify the look and feel of a painting and twenty different materials ranging from sky and mountains to river and stone. Paint on different layers to keep elements separate. You can start from scratch or get inspired by one of the sample scenes”.

In simpler words, perhaps, without having any particular drawing, painting, photographic or design skills, you use the mouse and apply rough, approximate brushstrokes to an empty window on your screen, select one among the above mentioned nine styles (mountain, sky, etc.), and Canvas almost instantly turns them into real places, objects, landscapes, as surely as if they were photographs of the actual place, object, or landscape. Once the target is created, you can customize, fine-tune, and modify it to your liking or to the requirements of the project you may be working on. It is done at lightning speed and the result is as realistic as if you took a photo — and a great one — of an actual place or landscape.

Dall-E-2 is the other application, and it works differently. No brushstrokes here. You just describe, with words, with plain text, the scenery you are thinking of, and Dall-E-2 will generate the photo you have in mind. For instance, type “a detached house, with brick walls, a swimming pool, a small red car in the garage, a dog in the garden, and cloudy skies above”, and watch the result on the screen, as photo realistic as in real life.

As with Canvas, watching pertaining videos on YouTube gives an accurate idea of the power of the application.

Customization, with painstaking details, is also possible here. Edit the text you started from, and the picture will follow, render, and apply the edit. You can even add people, flowers, or furniture. It is, again, the sheer power of software programming combined with AI algorithms that does the trick. It would have been unthinkable a mere 10 years ago.

Jordanian designer Yazan Baggili, who recently downloaded and experimented with Canvas in his office, explained that “Canvas is essentially about creating scenery and landscapes. Dalle-E-2 is more about visualizing things that are only in your mind, like for example: a cat in space in Van Gogh style!”
It is, again, the sheer power of software programming combined with AI algorithms that does the trick. It would have been unthinkable a mere 10 years ago.
Speaking to Jordan News he added: “We are slowly but surely getting used to how AI accesses gigantic data bases of just about any kind of content. Now Canvas and Dall-E-2 can tap these resources and create things that have never existed before, combining them in an artistic and astounding manner.”

To quickly name a third player in the game, Unreal 5 is another architectural visualization software that does incredible, photo realistic renderings.

It is understood that such applications require a fast computer processor, of the Intel i7 type, for example, and a powerful graphics processor as well, such as Nvidia of the latest generation. Not meeting these technical requirements may simply prevent the software from running at all or may even make installing such applications impossible.

We are barely starting to feel the effects of AI-injected software in a practical and tangible manner, whether in a highly positive way, or in a negative way, like in accidents and errors in predictions. In extreme cases, it can be tragic, at other times just entertaining.

A negative example: “US car manufacturers reported nearly 400 crashes involving cars with partially autonomous driver assistance systems, according to a new report from a US car-safety regulator released on Wednesday,” Al Jazeera published on June 15.

A lighter example is the amused reaction of Carlos Alija, the executive creative director at MullenLowe, a US advertising and marketing communications agency, at how tennis champion Rafael Nadal recovered from two sets down to win the Australian tennis open last January, when AI algorithms were giving his opponent Daniil Medvedev 96 percent chances to win during the 3rd set.

My first formal and academic approach to AI goes back to 1983, when I was studying the textbook titled “Artificial Intelligence” by Elaine Rich, from the University of Texas in Austin. Back then, it sounded like science fiction. It is solid reality today. As Canvas and Dall-E-2 are telling you, “see for yourself”.


Jean-Claude Elias is a computer engineer and a classically trained pianist and guitarist. He has been regularly writing IT articles, reviewing music albums, and covering concerts for more than 30 years.


Read more Opinion and Analysis
Jordan News
 
NEWS RELATED TO