Papers of the day   All papers

Vid2Game: Controllable Characters Extracted from Real-World Videos

Comments

Jonathan Fly πŸ‘Ύ: I am so excited for when @facebookai drops Vid2Game code. Imagine all the amazing things you could do with this. Maybe the models are even playable in a Colab? FMV games: The Next Generation? abs: https://arxiv.org/abs/1904.08379 vid: https://www.youtube.com/watch?v=GprHGpM11s0&feature=youtu.be https://t.co/usBwnae69p

7 replies, 411 likes


hardmaru: Vid2Game: Controllable Characters Extracted from Real-World Videos Given a video of a person performing an activity, they build a controllable model that can generate novel image sequences of that person, according to arbitrary user-defined controls. https://arxiv.org/abs/1904.08379 https://t.co/WielMV7JY4

0 replies, 154 likes


RenΓ© Schulte: AI to turn a human video into a game character! πŸŽ‰ Vid2Game is a new Deep Learning approach using multiple models to extract the human pose and mask and use that as a controllable game character. 😲 Video: https://www.youtube.com/watch?v=sNp6HskavBE Paper: https://arxiv.org/abs/1904.08379 #AI #AR #DNN https://t.co/s2eyUK67Kh

1 replies, 13 likes


Connor Shorten: Really cool! NBA 2k25? πŸ˜‚

0 replies, 8 likes


Swapnil Sawant: So interesting! Was just looking at pix2pix generated videos that generate realistic city navigation. Now we have game characters you can interact with in this generated environment.

1 replies, 5 likes


Content

Found on Dec 27 2019 at https://arxiv.org/pdf/1904.08379.pdf

PDF content of a computer science paper: Vid2Game: Controllable Characters Extracted from Real-World Videos