It's like a tech demo, a preview of the future. Give it 5 years and it will be super refined and probably the future of low cost animation for kids TV shows and stuff. Then even further, like how no one animates without a computer now, no one will animate without AI assistance.
Idk about the future of animation. It would require look-alike actors. Wouldn't it be easier to use already available tech to transfer gestures and facial expressions to cartoon models?
As far as I can tell, it really depends on the intensity of the "slider".
When the insensity of the style transfer is pushed mostly to the right (high), it just seems like Pixar or cartoons. Nothing uncanny whatsoever.
But when they show is about a quarter of the way to the right... it's utter nightmare fuel, like plastic surgery taken way too far. The worst kind of uncanny valley, so I definitely agree with you there.
Effectively these let an app (eg some VToonify tool) generate content that from the perspective of your live streaming app look like they are from a webcam
I'm glad things are progressing, but it bugs me that AI is largely being innovated for the use of... things like this? I know this comment is a bit disparaging and minimizes greater achievements, and I apologize for that, but the closeness of content-consumerism and AI is becoming quite off putting.
Based on the numbers in the paper this is just a little bit too slow for use as a real time video effect. At ~0.1 seconds per frame we just need about a 3x improvement in performance to get to 30fps “real time” video frame rates.
And on that thought since it appears they used nVidia hardware based on the CUDA dependency, it would be interesting to see how this performs on something like an M1/M2 where there’s dedicated ML hardware to help offload and accelerate things.
I missed the bit about using 8 of them to run it! Wow that’s a lot of GPU horsepower to do this. More efficient to just use a vtuber style pipeline using unreal engine and the metahumans or other avatars… only need one good GOU for that.
M1 performs better in some realtime use-cases because of the unified memory: the GPU and ML hardware can work on a camera framebuffer directly without any copy.
CUDA always requires sending data over the PCI bus, at least when it comes to realtime camera processing. GPUDirect exists but it's optimized for disks and NICs, I don't believe it's possible to use it with cameras.
No idea actually, I just find all sorts of odd benchmarks crop up for things where the Unified Memory Architecture on the M1/M2 give things surprising good performance due to the DMA transfer performance hit on other CPU/GPU combinations… it’s far from universal, but it’s just been surprising to see and this looked like the sort of thing that might be one of them between the camera decoding the ML & GPU processing and then “rendering” back out… where it might have had some benefits, hence my “wondering out loud”.
It’s been a while since I bothered looking at anything made by nVidia, I didn’t like their business practices and to be honest… Living in the Mac ecosystem for 90% of my work/life means I followed the paltry AMD rocM stuff ardently and was so happy when Apple made an effort up front to get the M1 and M2 chips supported by the major ML frameworks.
That’s a good point. I did briefly think about 24fps for “cinema” but as that’s a bit of a weird frame rate for a streaming webcam I dismissed the idea. But the video does actually have just enough frame rate to look smooth and I suppose that’s because it’s landing near enough to 12fps that combined with “looks like cartoon”, my brain is filing in the rest and barely noticing.
