Nvidia's Outrageous AI Solution for the Video
Conferencing Era
Nvidia develops Vid2Vid Cameo, which creates a Talking Head composite video as if the participant is actually speaking only with 2D images through GAN.
If you upload your own photo or character avatar, the AI model captures the user's movement in real time during the video conference and applies it to the submitted image, Based on this photo, neat hair and wearing a suit are mapped neatly, so even when you wake up, video conference with the perfect appearance.
Source: Nvidia
👉• Web: http://www.nvidia.com
👉• Paper: https://arxiv.org/pdf/2011.15126.pdf
👉• Video: NVIDIA - YouTube
0 Comments