Research map / Label-efficient and robust learning / Noisy labels
Collaborative Noisy Label Cleaner: Learning Scene-Aware Trailers for Multi-Modal Highlight Detection in Movies
Bei Gan, Xiujun Shu, Ruizhi Qiao, Haoqian Wu, Keyu Chen, Hanjun Li, Bo Ren
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 CLC instead of UMT | Movie highlight detection · MovieLights | mAP: 38.7 -> 43.88 | Improved |
| Use CLC instead of UMT | YouTube Highlights · YouTube Highlights | average mAP: 74.9 -> 80.1 | Improved |
| Use CLC instead of UMT under noisy training labels | YouTube Highlights · Harvested matched (clean) | average mAP: 74.90 -> 80.05 (↑5.15) | Improved |
| Use CLC instead of UMT under slight label noise | YouTube Highlights · Harvested borderline | average mAP: 74.39 -> 79.71 (↑5.32) | Improved |
| Use CLC instead of UMT under severe label noise | YouTube Highlights · Mturk | average mAP: 73.36 -> 79.22 (↑5.86) | Improved |
| Add multi-modality sample cleaning (ablation) | CLC- · MovieLights | mAP: 39.65 -> 41.69 | Improved |
| Add cross-propagation (ablation) | CLC with multi-modality sample cleaning · MovieLights | mAP: 41.69 -> 42.79 | Improved |
| Add consistency loss (ablation) | CLC with multi-modality sample cleaning and cross-propagation · MovieLights | mAP: 42.79 -> 43.22 | Improved |
| Add median-filter post-processing (ablation) | CLC with multi-modality sample cleaning, cross-propagation, and consistency loss · MovieLights | mAP: 43.22 -> 43.88 | Improved |
Closest papers by meaning
- TVHighlights: LLM-Guided Human-Free Collaborative Training for Video Highlight Detection in Movies and TV Dramas (CVPR 2026 · similarity 0.91)
- Joint Visual and Audio Learning for Video Highlight Detection (ICCV 2021 · similarity 0.87)
- Temporal Cue Guided Video Highlight Detection With Low-Rank Audio-Visual Fusion (ICCV 2021 · similarity 0.87)
- Cross-Category Video Highlight Detection via Set-Based Learning (ICCV 2021 · similarity 0.87)
- Learning to Highlight Audio by Watching Movies (CVPR 2025 · similarity 0.86)
- CineTechBench: A Benchmark for Cinematographic Technique Understanding and Generation (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.
