In the image, a man is seen attending a conference presentation from the back of a conference room. A large screen is visible behind him, displaying a series of charts and graphs related to the subject matter being discussed. The man is seated by a window, with a laptop open in front of him, possibly taking notes or working on his own project. To his right, a cup is placed on a table, and a television is visible on the other side of the room, displaying a PowerPoint presentation that appears to be the main focus of the conference. The room appears to be in a professional setting, with a board room table and chairs arranged in a semi-circle, suggesting a collaborative and interactive atmosphere.
Text transcribed from the image:
Lumiere
A space-time diffusion model for video generation
(a) Space-Time UNet (STUNet)
W
TX H X W X D
Legend:
00
Spatial Resizing
A Temporal Resizing
Skip Connection
Conv-based Inflation
Attention-based Inflation
글x블x쁠x ++
l et al., "Lumiere: A space-time diffusion model for video generation," arXiv 2024.
(b) Convolution-based Inflation Block
Pretrained Spatial Layer(s)
2D Convolution
Norm+ activiation
ID Convolution
Norm+acaviation
Linear Projection
(c) Attention-based Inflation
Pretrained Spati
IDA
Copyright
Linca