AI Paper Atlas

Research map / Video / Video-language understanding

StreamRAG: Enhancing Real-Time Video Understanding with Retrieval Augmentation

Junlin Xie, Quanlong Zheng, Ruifei Zhang, Kuo Wang, Yanhao Zhang, Jinguo Luo, Haonan Lu, Xiang Wan, Guanbin Li

CVPR 2026 · Video › Video-language understanding · approach: Long video understanding

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add Knowledge Extraction Accelerator (retrieval knowledge base) (ablation)Qwen2-VL-7B · OVO-BenchR-Avg: 55.98 -> 65.33Improved
add Streaming Event Segmentation on top of KEA (ablation)StreamRAG · OVO-BenchR-Avg: 65.33 -> 68.54Improved
add Dynamic Knowledge Integration on top of KEA+SES (ablation)StreamRAG · OVO-BenchB-Avg: 51.28 -> 53.08Improved
reuse about 18% of caption tokens (ablation)Knowledge Extraction Accelerator · OVO-Benchlatency: 15.24s -> 11.12sImproved
increase token reuse from about 18% to about 33% (ablation)Knowledge Extraction Accelerator · OVO-BenchB-Avg: 53.08 -> 50.47Hurt
increase token reuse from about 18% to about 33% (ablation)Knowledge Extraction Accelerator · OVO-Benchlatency: 11.12s -> 8.43sImproved
use event-triggered segmentation instead of fixed-interval updates (ablation)StreamRAG · OVO-BenchB-Avg: 50.04 -> 53.08Improved
use more frames for fixed-interval updating (16 to 32) (ablation)Fixed-Interval baseline · OVO-BenchR-Avg: 65.33 -> 68.98Improved
adaptive retrieval scope instead of latest-only caption (ablation)StreamRAG · OVO-BenchB-Avg: 51.25 -> 53.08Improved
adaptive retrieval scope instead of using the full knowledge base (ablation)StreamRAG · OVO-BenchB-Avg: 51.28 -> 53.08Improved
add offline VideoRAG to ViSpeakViSpeak · OVO-Bench Real-Timeavg accuracy: 66.28 -> 57.21Hurt
add StreamChat-style RAG with uniform sampling to ViSpeakViSpeak · OVO-Bench Real-Timeavg accuracy: 66.28 -> 64.48Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy