TLDR Tech / AI
Interaction-Aware Video Object Removal (GitHub Repo)
Netflix's VOID is a video inpainting model that removes objects along with their physical and visual interactions, using interaction-aware mask conditioning built on CogVideoX.
Read the full story