Generative Animation and the Future of Interactivity

Lately, more and more materials are appearing online where static images start moving, dancing, and jumping. It all looks pretty odd, and at this point, it might not be entirely clear why this is needed. In this small article, I would like to dive into this topic and talk a little about where these "funny" mini-videos on Twitter and other social networks might lead.

As we know, everything starts small, so in our case, these demonstrations are just the beginning of something big. But what exactly — let's try to figure it out!

Let's dive a bit into history, because as they say, "Who knows history, knows the future." Remember 2010, when computer vision was just beginning to develop, and we could use primitive apps. But it was only in 2015 that Facebook acquired the startup MSQRD, which allowed users to apply various filters and masks to their faces in real-time, tracking our faces. No one could imagine that in about 5-6 years, companies would start creating digital avatars and use them instead of real images from webcams.

Operating with these facts allows us to assume that in this way, in about 3 years, Moore's Law will significantly influence strategic planning and expectations in many technological fields, pushing the industry towards innovation and continuous improvement of technology performance. We might be surprised at what these dancing figures could transform into.

Considering the example of Persona (a technology from Apple that allows creating a photorealistic avatar and using it during video calls), we can imagine that in 5 years, they will find a way for avatars to appear in front of you, walk with you, interact. Just like in the most daring fantasy movies. This will be possible thanks to computer vision on the one hand, and on the other, algorithms that can transform 2D photos into 3D movements.

From my own experience in developing XR apps and games, I can say that rigging animations is such a time-consuming task that all the hours in the world of all designers won't allow us to work out animations that would cover all interaction cases and give us a feeling of interacting with a live character or avatar. Therefore, the future is for generative animation. I will write more about this in another article, but for now, let's continue with our topic.

Another example of use is showing clothing on models in video. Because in those same 3 years, I am sure that this technology will be actively used in production. Firstly, it's much cheaper than launching a real production, and secondly, you can create much more diverse content.

Fitness apps are a whole industry where videos from real gyms, real trainers are often used now. Again, with the use of movement generation, this expands the possibilities for designers and developers of such apps. The same quality level can be achieved with less effort and a smaller budget.

As you can see, at first glance, clumsy demonstrations that we see now can become very promising technologies in the future. Therefore, as practice shows, R&D projects usually turn into something bigger. Thus, if companies have the capabilities and resources, it's worth engaging in R&D, as it helps make our world better!