Near-perfect facial animations made with Unreal Engine 5!

Announced by game engine maker Epic Games on May 13, 2020, and released into early access last May. Unreal Engine 5 He enchanted the audience with his first demos. However, this week Ziva Dynamics The new videos published by the developer named , confronted us with the best face animations ever.


Unreal Engine 5 demo character coming to Fortnite

A character from the tech demo shown on PS5 a long time ago is coming to Fortnite. The character will be available tonight.

Unreal 5 takes visuals to another level

As you may remember, in May of last year, the announcement of Unreal Engine 5 was announced with a real-time demo on PlayStation 5. The demo, which fascinates the audience with its interactivity, dynamic lighting and incredible level of detail, what we might face in the future was revealing. It was even claimed that the published video was prepared in advance and was CGI.

SDNshiftdelete.net

While contemplating how far the demo, which you can watch again above, can be taken, the developer named Ziva Dynamics has announced its brand new cloud-based face animation browser. ZRT Face Trainer In its beta version, the project uses machine learning technique to create the most realistic facial animations we have ever seen.

The video posted on Ziva’s YouTube channel, which you can watch below, was shot in real time at 4K and 60FPS. ZRT Face Trainer’s database, 15TB of 4D images includes and in total 72,000 different facial expressions It can emulate in real time with Unreal Engine 5.

SDNshiftdelete.net

Finally, we should add: this realized project may never reach the end user. Instead, the company wants to use this technology in their games. to various game makers Will try marketing. However, if the technology they show is already at this level, we can watch unbelievable animations in the games we will play in the near future.

So what do you think about Unreal Engine 5 and Ziva RealTime Face Trainer? Do not forget to share your views with us in the comments section.

source site