Event-VStream: Event-Driven Real-Time Understanding for Long Video Streams
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866912840499068928 |
|---|---|
| author | Guo, Zhenghui Man, Yuanbin Sheng, Junyuan Lin, Bowen Ahmed, Ahmed Jiang, Bo Zhang, Boyuan Yin, Miao Jin, Sian Gnawal, Omprakash Zhang, Chengming |
| author_facet | Guo, Zhenghui Man, Yuanbin Sheng, Junyuan Lin, Bowen Ahmed, Ahmed Jiang, Bo Zhang, Boyuan Yin, Miao Jin, Sian Gnawal, Omprakash Zhang, Chengming |
| contents | Real-time understanding of long video streams remains challenging for multimodal large language models (VLMs) due to redundant frame processing and rapid forgetting of past context. Existing streaming systems rely on fixed-interval decoding or cache pruning, which either produce repetitive outputs or discard crucial temporal information. We introduce Event-VStream, an event-aware framework that represents continuous video as a sequence of discrete, semantically coherent events. Our system detects meaningful state transitions by integrating motion, semantic, and predictive cues, and triggers language generation only at those boundaries. Each event embedding is consolidated into a persistent memory bank, enabling long-horizon reasoning while maintaining low latency. Across OVOBench-Realtime, and long-form Ego4D evaluations, Event-VStream achieves competitive performance. It improves over a VideoLLM-Online-8B baseline by +10.4 points on OVOBench-Realtime, achieves performance close to Flash-VStream-7B despite using only a general-purpose LLaMA-3-8B text backbone, and maintains around 70% GPT-5 win rate on 2-hour Ego4D streams. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_15655 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Event-VStream: Event-Driven Real-Time Understanding for Long Video Streams Guo, Zhenghui Man, Yuanbin Sheng, Junyuan Lin, Bowen Ahmed, Ahmed Jiang, Bo Zhang, Boyuan Yin, Miao Jin, Sian Gnawal, Omprakash Zhang, Chengming Computer Vision and Pattern Recognition Artificial Intelligence Real-time understanding of long video streams remains challenging for multimodal large language models (VLMs) due to redundant frame processing and rapid forgetting of past context. Existing streaming systems rely on fixed-interval decoding or cache pruning, which either produce repetitive outputs or discard crucial temporal information. We introduce Event-VStream, an event-aware framework that represents continuous video as a sequence of discrete, semantically coherent events. Our system detects meaningful state transitions by integrating motion, semantic, and predictive cues, and triggers language generation only at those boundaries. Each event embedding is consolidated into a persistent memory bank, enabling long-horizon reasoning while maintaining low latency. Across OVOBench-Realtime, and long-form Ego4D evaluations, Event-VStream achieves competitive performance. It improves over a VideoLLM-Online-8B baseline by +10.4 points on OVOBench-Realtime, achieves performance close to Flash-VStream-7B despite using only a general-purpose LLaMA-3-8B text backbone, and maintains around 70% GPT-5 win rate on 2-hour Ego4D streams. |
| title | Event-VStream: Event-Driven Real-Time Understanding for Long Video Streams |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2601.15655 |