NVIDIA broadcast receives an update that opens the Eye Contact feature to the public. This allows NVIDIA RTX users to appear as if they are looking directly into the camera’s lens, when in fact they are looking in a slightly different direction.
If you want to look directly at the camera while recording video, you generally need a teleprompter-like setup, and that’s not practical for ordinary users. NVIDIA Broadcast is software that uses AI (artificial intelligence) to modify video input and change the direction of your gaze.
Looking straight into the camera is more visually appealing to your audience, so this feature could add a lot of value, especially since it’s very handy and easy to use (you just have to turn it on).
And it works surprisingly well when looking in the general direction of the camera, which I estimate is within 35 degrees of the lens. However, it would be best if you did your best to face the camera directly for the best result.
At this point there are still minor quirks. For example, if you blink, the AI may lose sight of your gaze for a second or two, causing your eyes to abruptly change direction. The same could happen if you turn your head away too far and come back.
NVIDIA may be able to fix this later by remembering where your staring eye was before you blinked. Other than that, it works well enough and so far no one has noticed that I’ve used it during a call. At least not in the sense of commenting on it.
NVIDIA Broadcast and its features are essentially the result of what the company has been doing with its NVIDIA AI Research group. Many webcam-related features were demonstrated as research projects a few years or more ago, and it’s nice to see a steady stream of improvements to the publicly available free software. It is also open to developers.
Hopefully NVIDIA will release its AI Face Alignment feature at some point because either would make this look even more natural. While facial realignment is one thing, doing it to the face and torso is another. I suspect there is one more challenge to overcome before this feature is released to everyone.
This type of workload shows that GPUs are useful to a much larger group of users than 3D graphics for gaming or CAD.
Filed in . Read more about AI (Artificial Intelligence) and NVIDIA.
This article was previously published on Source link