Research map / Video / Video segmentation and tracking
A Deep Ranking Model for Spatio-Temporal Highlight Detection From a 360◦ Video
Youngjae Yu, Sangho Lee, Joonil Na, Jaeyun Kang, Gunhee Kim
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 smooth-motion stitching to CVS-C3D (ablation) | CVS-C3D · Pano2Vid | Frame cosine similarity: 0.656 -> 0.774 | Improved |
| Add smooth-motion stitching to CVS-C3D (ablation) | CVS-C3D · Pano2Vid | Frame overlap: 0.554 -> 0.646 | Improved |
| Add smooth-motion stitching to CVS-Inception (ablation) | CVS-Inception · Pano2Vid | Frame cosine similarity: 0.642 -> 0.768 | Improved |
| Add smooth-motion stitching to CVS-Inception (ablation) | CVS-Inception · Pano2Vid | Frame overlap: 0.545 -> 0.666 | Improved |
| Add smooth-motion stitching to CVS-Fusion (ablation) | CVS-Fusion · Pano2Vid | Frame cosine similarity: 0.701 -> 0.800 | Improved |
| Add smooth-motion stitching to CVS-Fusion (ablation) | CVS-Fusion · Pano2Vid | Frame overlap: 0.590 -> 0.677 | Improved |
| Fuse C3D and Inception-v3 features rather than using C3D alone (ablation) | CVS · Pano2Vid | Frame cosine similarity: 0.656 -> 0.701 | Improved |
| Fuse C3D and Inception-v3 features rather than using Inception-v3 alone (ablation) | CVS · Pano2Vid | Frame overlap: 0.545 -> 0.590 | Improved |
| Use triplet ranking loss instead of pairwise ranking loss (ablation) | CVS-Fusion · Wedding | mAP: 14.34 -> 17.96; 3.62 | Improved |
| Use triplet ranking loss instead of pairwise ranking loss (ablation) | CVS-Fusion · MV | mAP: 12.56 -> 14.92; 2.36 | Improved |
| Use triplet ranking loss rather than pairwise ranking models (ablation) | CVS-Fusion · Wedding | mAP: +5.98 vs RankNet; +4.73 vs TS-DCNN | Improved |
| Use triplet ranking loss rather than pairwise ranking models (ablation) | CVS-Fusion · MV | mAP: +3.27 vs RankNet; +2.64 vs TS-DCNN | Improved |
Closest papers by meaning
- Mr. HiSum: A Large-scale Dataset for Video Highlight Detection and Summarization (NeurIPS 2023 · similarity 0.90)
- Exploring Domain Incremental Video Highlights Detection with the LiveFood Benchmark (AAAI 2024 · similarity 0.86)
- Bridging the Gap: A Unified Video Comprehension Framework for Moment Retrieval and Highlight Detection (CVPR 2024 · similarity 0.86)
- A Motion-aware Spatio-temporal Graph for Video Salient Object Ranking (NeurIPS 2024 · similarity 0.85)
- Watch Video, Catch Keyword: Context-aware Keyword Attention for Moment Retrieval and Highlight Detection (AAAI 2025 · similarity 0.85)
- CSTA: CNN-based Spatiotemporal Attention for Video Summarization (CVPR 2024 · 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.
