CamI2V-Epipolar: Epipolar-Constrained Block Sparse Attention for Camera-Controlled Image-to-Video Diffusion Model
IEEE Signal Processing Letters · 2026 · 33 · 1401-1405
DOI: 10.1109/lsp.2026.3675911 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Signal Processing Letters, 2026. Verfasst von Guangcong Zheng und 1 weiteren Autor:innen. Die Arbeit nennt 43 Literaturangaben.
Abstract (Originalsprache)
Recently, camera pose has emerged as a physics-informed condition for video diffusion models. Existing methods that directly adopt 3D full attention for cross-frame feature interaction achieve moderate camera controllability with extensive training, yet camera controllability and geometric consistency remain a challenge in complex scenarios such as large camera rotation or movement. This underscores the value of integrating physical priors into attention mechanisms. Instead of pixel-level attention masking, we propose a key innovation to infuse epipolar geometric priors into the block selection of video block sparse attention. This design resolves memory and speed bottlenecks at high resolutions like 768P, enables compatibility with FlashAttention-3, and establishes a principled framework for embedding physical priors into sparse attention mechanisms. Experiments on static RealEstate10 K and dynamic RealCam-Vid datasets demonstrate that our method outperforms the state-of-the-art RealCam-I2V, effectively enhancing camera controllability while preserving generation quality and generalization.
Zitieren
APA
Zheng, G., & Li, X. (2026). CamI2V-Epipolar: Epipolar-Constrained Block Sparse Attention for Camera-Controlled Image-to-Video Diffusion Model. IEEE Signal Processing Letters, 33, 1401-1405. https://doi.org/10.1109/lsp.2026.3675911
Vancouver
Zheng G, Li X. CamI2V-Epipolar: Epipolar-Constrained Block Sparse Attention for Camera-Controlled Image-to-Video Diffusion Model. IEEE Signal Processing Letters. 2026;33:1401-1405. doi:10.1109/lsp.2026.3675911
BibTeX
@article{Zheng2026,
author = {Zheng, Guangcong and Li, Xi},
title = {CamI2V-Epipolar: Epipolar-Constrained Block Sparse Attention for Camera-Controlled Image-to-Video Diffusion Model},
journal = {IEEE Signal Processing Letters},
year = {2026},
volume = {33},
pages = {1401--1405},
doi = {10.1109/lsp.2026.3675911}
}
EUVIMED