Research map / Video / Video-language understanding
TripleSumm: Adaptive Triple-Modality Fusion for Video Summarization
Sumin Kim, Hyemin Jeong, Mingu Kang, Yejin Kim, Yoori Oh, Joonseok Lee
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add text and audio to visual input (trimodal) (ablation) | TripleSumm · MoSu | Kendall tau: 0.298 -> 0.351 | Improved |
| add multi-scale temporal block to cross-modal fusion only (ablation) | TripleSumm · MoSu | Kendall tau: 0.250 -> 0.351 | Improved |
| add cross-modal fusion block to MST only (ablation) | TripleSumm · MoSu | Kendall tau: 0.337 -> 0.351 | Improved |
| use dynamic per-frame fusion instead of static averaging (ablation) | TripleSumm · MoSu | Kendall tau: 0.338 -> 0.351 | Improved |
| use dynamic per-frame fusion instead of global weights (ablation) | TripleSumm · MoSu | Kendall tau: 0.344 -> 0.351 | Improved |
| use local-to-global window schedule instead of constant global window (ablation) | TripleSumm · MoSu | Kendall tau: 0.317 -> 0.351 | Improved |
| use local-to-global instead of global-to-local window schedule (ablation) | TripleSumm · MoSu | Kendall tau: 0.345 -> 0.351 | Improved |
| share temporal block parameters across modalities (ablation) | TripleSumm · MoSu | Kendall tau: 0.343 -> 0.351; params 2.95M -> 1.37M | Improved |
| use average of modalities as fusion token instead of learnable (ablation) | TripleSumm · MoSu | Kendall tau: 0.341 -> 0.351 | Improved |
| use average fusion token instead of visual-only token (ablation) | TripleSumm · MoSu | Kendall tau: 0.337 -> 0.351 | Improved |
| add learnable Gaussian window modulation (ablation) | TripleSumm · MoSu | Kendall tau: 0.351 -> 0.353 | No clear effect |
| increase layers from 2 to 8 (ablation) | TripleSumm · MoSu | Kendall tau: 0.351 -> 0.326 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CSTA — CSTA: CNN-based Spatiotemporal Attention for Video Summarization (CVPR 2024) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- V2Xum-LLM: Cross-Modal Video Summarization with Temporal Prompt Instruction Tuning (AAAI 2025 · similarity 0.89)
- Video Summarization with Large Language Models (CVPR 2025 · similarity 0.89)
- MMSum: A Dataset for Multimodal Summarization and Thumbnail Generation of Videos (CVPR 2024 · similarity 0.88)
- Video-MME: The First-Ever Comprehensive Evaluation Benchmark of Multi-modal LLMs in Video Analysis (CVPR 2025 · similarity 0.87)
- SF2T: Self-supervised Fragment Finetuning of Video-LLMs for Fine-Grained Understanding (CVPR 2025 · similarity 0.86)
- MotionBench: Benchmarking and Improving Fine-grained Video Motion Understanding for Vision Language Models (CVPR 2025 · similarity 0.85)
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.
