Research map / Video / Video segmentation and tracking
Movies2Scenes: Using Movie Metadata To Learn Scene Representation
Shixing Chen, Chun-Hao Liu, Xiang Hao, Xiaohan Nie, Maxim Arap, Raffay Hamid
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 |
|---|---|---|---|
| Use Movies2Scenes scene-contrast representations instead of SlowFast representations | LVU place retrieval · LVU place retrieval, all queries | top-1 precision: 38.59 -> 47.36 | Improved |
| Use Movies2Scenes scene-contrast representations instead of CLIP representations | LVU place retrieval · LVU place retrieval, all queries | top-5 precision: 29.64 -> 39.29 | Improved |
| Use Movies2Scenes scene-contrast representations instead of CLIP representations | LVU place retrieval · LVU place retrieval, all queries | top-10 precision: 28.85 -> 35.70 | Improved |
| Use Movies2Scenes pre-training instead of MerlotReserve pre-training | MCD age-appropriate activity classification · MCD | average: 66.2 -> 71.1 | Improved |
| Use Movies2Scenes pre-training instead of MerlotReserve pre-training | MCD age-appropriate activity classification · MCD | sex: 77.0 -> 81.5 | Improved |
| Use Movies2Scenes pre-training instead of MerlotReserve pre-training | MCD age-appropriate activity classification · MCD | violence: 68.2 -> 70.2 | Improved |
| Use Movies2Scenes pre-training instead of BridgeFormer pre-training | MCD age-appropriate activity classification · MCD | drug-use: 61.2 -> 61.8 | Improved |
Later papers that compared with this method
1 later accepted paper reports Movies2Scenes as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- HERMES: temporal-coHERent long-forM understanding with Episodes and Semantics (ICCV 2025 · better on 1 of 1 shared result)
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
- A Local-to-Global Approach to Multi-Modal Movie Scene Segmentation (CVPR 2020 · similarity 0.88)
- MovieNet: A Holistic Dataset for Movie Understanding (ECCV 2020 · similarity 0.88)
- Video Scene Segmentation with Genre and Duration Signals (ICLR 2026 · similarity 0.87)
- Learning Trailer Moments in Full-Length Movies with Co-Contrastive Attention (ECCV 2020 · similarity 0.86)
- MovieBench: A Hierarchical Movie Level Dataset for Long Video Generation (CVPR 2025 · similarity 0.85)
- ConViS-Bench: Estimating Video Similarity Through Semantic Concepts (NeurIPS 2025 · similarity 0.84)
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.
