EM-Vid: Training-Free Entity-Centric Memory for Efficient and Consistent Multi-Shot Video Generation

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Vandersanden, Jente, Gadelha, Matheus, Huang, Chun-Hao P., Jeong, Hyeonho, Gryaditskaya, Yulia
Format:	Preprint
Published:	2026
Subjects:	Computer Vision and Pattern Recognition Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2605.23610