If you are looking for a of a specific section or want to know how to run the code , let me know!
Paper2Video: Automatic Video Generation from Scientific Papers
The paper introduces , a multi-agent framework that automatically converts academic papers into professional presentation videos. It breaks the process down into four distinct "builders": 0h7c8bggs3o0hh72h4fi4_source.mp4
: Synchronizes a virtual cursor with the narration to highlight specific areas of the slides.
: Generates a realistic, personalized talking-head video using a portrait and voice sample. If you are looking for a of a
You can find more details, the full paper, and video demos on the official Paper2Video Project Page .
: Uses Vision-Language Models (VLMs) to create narration subtitles and visual-focus prompts. : Generates a realistic
: Automatically generates and refines LaTeX-based slides from the paper's text.