UniScene-MoTion: Unified Scene & Motion-aware Diffusion Transition Framework
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (7) · 5415-5423
DOI: 10.1609/aaai.v40i7.37458 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Rui Jiang und 5 weiteren Autor:innen.
Abstract (Originalsprache)
Video transitions are critical for ensuring temporal coherence in edited media, yet existing methods often rely on handcrafted effects or relative-scale trajectories that fail to capture the physical structure of real-world scenes. In this work, we introduce a scale-aware video transition framework that explicitly incorporates depth-aware 3D reasoning into a diffusion-based generation pipeline. Built upon a powerful I2V foundation, our method leverages single-image depth prediction to align camera motion with metric-scale geometry, enabling physically consistent transitions. To reduce reliance on precise camera inputs, we propose a bidirectional conditional control module and a progressive training strategy with conditional dropout, enhancing generalization to loosely specified or missing camera trajectories. Extensive experiments demonstrate that our approach achieves state-of-the-art performance, delivering realistic, geometrically coherent transitions across diverse scenes and applications with minimal input guidance.
Zitieren
APA
Jiang, R., Wang, C., Fu, X., Lu, Y., Li, T., & Li, X. (2026). UniScene-MoTion: Unified Scene & Motion-aware Diffusion Transition Framework. Proceedings of the AAAI Conference on Artificial Intelligence, 40(7), 5415-5423. https://doi.org/10.1609/aaai.v40i7.37458
Vancouver
Jiang R, Wang C, Fu X, Lu Y, Li T, Li X. UniScene-MoTion: Unified Scene & Motion-aware Diffusion Transition Framework. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(7):5415-5423. doi:10.1609/aaai.v40i7.37458
BibTeX
@article{Jiang2026,
author = {Jiang, Rui and Wang, Chongmian and Fu, Xinghe and Lu, Yehao and Li, Teng and Li, Xi},
title = {UniScene-MoTion: Unified Scene & Motion-aware Diffusion Transition Framework},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {7},
pages = {5415--5423},
doi = {10.1609/aaai.v40i7.37458}
}
EUVIMED