Presentation

<
LongE2V: Long-Horizon Event-based Video Reconstruction, Prediction, and Frame Interpolation with Video Diffusion Models
SessionFilm/VFX Production
DescriptionThis work introduces long-horizon event-based video generation, a challenging task prone to severe temporal inconsistency, and presents LongE2V, a novel approach leveraging video diffusion priors to jointly handle video reconstruction, prediction, and frame interpolation, recovering high-quality videos from sparse event streams.
Authors
Cheng-De Fan
National Yang Ming Chiao Tung University
Chun-Wei Tuan Mu
National Yang Ming Chiao Tung University
Chen-Wei Chang
National Yang Ming Chiao Tung University
Chin-Yang Lin
National Yang Ming Chiao Tung University
Kun-Ru Wu
National Yang Ming Chiao Tung University
Yu-Chee Tseng
National Yang Ming Chiao Tung University
Yu-Lun Liu
National Yang Ming Chiao Tung University
Event Type
Technical Paper
TimeWednesday, 22 July 20262:30pm - 2:40pm PDT
LocationRoom 408 B
Digital Library PDF
Session Time & Location
Sunday, 19 July 20266:00pm - 8:45pm PDTHall K
Wednesday, 22 July 20262:00pm - 3:30pm PDTRoom 408 B
Keywords
Artificial Intelligence/Machine Learning
Generative AI
Registration Categories
Full Conference Supporter
Full Conference
Next PresentationNext PresentationAutoregressive Modeling of Film with Applications in Video Montage
Similar Presentations
Just-Dub-It: Video dubbing via Joint Audio-Visual Diffusion
See-through: Single-image Layer Decomposition for Anime Characters
BodyReLux: Temporally Consistent Full-Body Video Relighting