In this image, a group of people are attending a presentation in a conference room. The main focus is a large screen at the front displaying a slide titled "Text2Video-Zero." The slide discusses using Stable Diffusion to generate videos without fine-tuning and includes diagrams and text for deeper explanation. Attendees are seated and engaged with the material, some taking notes. The room is equipped with modern lighting and ceiling panels and has a clean, minimalistic design. On the floor, there is a yellow cloth and some personal items near the front row, indicating someone might have stepped away momentarily. The atmosphere in the room suggests a professional or academic environment focused on technology or research.
Text transcribed from the image:
Text2Video-Zero
Use Stable Diffusion to generate videos without any finetuning
• Optional background smoothing: regenerate the background, average with the first frame
-N(0,1)
-DOO
DOPM Fr
Background Smoothing
Dachatryan et al, "exvides-Zera Test to-image Diffusion Models are Zero-Shot Video Generators," anov 2023.
Copyright CMike Show, NUS