Video Echoed in Music: Semantic, Temporal, and Rhythmic Alignment for Video-to-Music Generation
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (31) · 25983-25991
DOI: 10.1609/aaai.v40i31.39799 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Xinyi Tong und 11 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref).
Abstract (Originalsprache)
Video-to-Music generation seeks to generate musically appropriate background music that enhances audiovisual immersion for videos. However, current approaches suffer from two critical limitations: 1) incomplete representation of video details, leading to weak alignment, and 2) inadequate temporal and rhythmic correspondence, particularly in achieving precise beat synchronization. To address the challenges, we propose Video Echoed in Music (VeM), a latent music diffusion that generates high-quality soundtracks with semantic, temporal, and rhythmic alignment for input videos. To capture video details comprehensively, VeM employs a hierarchical video parsing that acts as a music conductor, orchestrating multi-level information across modalities. Modality-specific encoders, coupled with a storyboard-guided cross-attention mechanism (SG-CAtt), integrate semantic cues while maintaining temporal coherence through position and duration encoding. For rhythmic precision, the frame-level transition-beat aligner and adapter (TB-As) dynamically synchronize visual scene transitions with music beats. We further contribute a novel video-music paired dataset sourced from e-commerce advertisements and video-sharing platforms, which imposes stricter transition-beat synchronization requirements. Meanwhile, we introduce novel metrics tailored to the task. Experimental results demonstrate superiority, particularly in semantic relevance and rhythmic precision.
Zitieren
APA
Tong, X., Zhu, Y., Chen, J., Zhan, C., Wang, T., Zhang, S., Liu, N., Ge, T., XU, D., Jin, X., Yu, F., & Zhu, S. (2026). Video Echoed in Music: Semantic, Temporal, and Rhythmic Alignment for Video-to-Music Generation. Proceedings of the AAAI Conference on Artificial Intelligence, 40(31), 25983-25991. https://doi.org/10.1609/aaai.v40i31.39799
Vancouver
Tong X, Zhu Y, Chen J, Zhan C, Wang T, Zhang S, et al. Video Echoed in Music: Semantic, Temporal, and Rhythmic Alignment for Video-to-Music Generation. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(31):25983-25991. doi:10.1609/aaai.v40i31.39799
BibTeX
@article{Tong2026,
author = {Tong, Xinyi and Zhu, Yiran and Chen, Jishang and Zhan, Chunru and Wang, Tianle and Zhang, Sirui and Liu, Nian and Ge, Tiezheng and XU, Duo and Jin, Xin and Yu, Feng and Zhu, Song-Chun},
title = {Video Echoed in Music: Semantic, Temporal, and Rhythmic Alignment for Video-to-Music Generation},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {31},
pages = {25983--25991},
doi = {10.1609/aaai.v40i31.39799}
}
EUVIMED