PreFM: Online Audio-Visual Event Parsing via Predictive Future Modeling
机构 * Institute of Information Science, Beijing Jiaotong University(信息科学学院,北京交通大学)
专题命中 动作与事件理解 :video understanding(abstract);分类 cs.CV
Comments This paper is accepted by 39th Conference on Neural Information Processing Systems (NeurIPS 2025)