MultiSoundGen: Video-to-Audio Generation for Multi-Event Scenarios via SlowFast Contrastive Audio-Visual Pretraining and Direct Preference Optimization.

@inproceedings{yang-2026-multisoundgen, author = {Jianxuan Yang and Xiaoran Yang and Lipan Zhang and Xinyue Guo and Zhao Wang and Gongping Huang}, editor = {Stevan Rudinac and Zi Huang and Marcel Worring and Lucia Vadicamo and Xirong Li and Pascal Mettes and Lizi Liao}, title = {{MultiSoundGen: Video-to-Audio Generation for Multi-Event Scenarios via SlowFast Contrastive Audio-Visual Pretraining and Direct Preference Optimization}}, booktitle = {{Proceedings of the 2026 International Conference on Multimedia Retrieval, ICMR 2026, Amsterdam, The Netherlands, June 16-19, 2026}}, pages = {1918--1926}, publisher = {ACM}, year = {2026}, url = {https://doi.org/10.1145/3805622.3810680}, doi = {https://doi.org/10.1145/3805622.3810680} }