Research map / Video / Video segmentation and tracking
Scene Consistency Representation Learning for Video Scene Segmentation
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 scene-consistency positive selection instead of nearest-neighbor selection (ablation) | SCRL · MovieNet | AP: +2.95 AP | Helped |
| Use self-augmented positive selection instead of scene-consistency selection (ablation) | SCRL · MovieNet | AP | Hurt |
| Add scene-agnostic clip-shuffling with NN selection (ablation) | SCRL with NN selection · MovieNet | AP: +1.85 AP | Helped |
| Add scene-agnostic clip-shuffling with scene-consistency selection (ablation) | SCRL with scene-consistency selection · MovieNet | AP: +2.46 AP | Helped |
| Use Soft-SC rather than multiple positive samples for scene consistency (ablation) | SCRL · MovieNet | AP: 53.74 AP | Helped |
| Use SimSiam instead of MoCo as the SSL framework (ablation) | SCRL · MovieNet | AP: 53.74 -> 39.82 | Hurt |
| Use SimCLR instead of MoCo as the SSL framework (ablation) | SCRL · MovieNet | AP: 53.74 -> 45.32 | Hurt |
| Use BYOL instead of SimSiam as the SSL framework (ablation) | SCRL · MovieNet | AP: 39.82 -> 50.35 | Helped |
| Increase MLP shot length from 4 to 10 (ablation) | SCRL with MLP evaluation · MovieNet | AP: 53.74 -> 49.61 | Hurt |
| Increase MLP shot length from 4 to 10 (ablation) | SCRL with MLP evaluation · MovieNet | F1: 50.40 -> 44.04 | Hurt |
| Increase Bi-LSTM shot length from 10 to 40 (ablation) | SCRL with Bi-LSTM evaluation · MovieNet | AP: 43.94 -> 54.55 | Helped |
| Increase Bi-LSTM shot length from 10 to 40 (ablation) | SCRL with Bi-LSTM evaluation · MovieNet | F1: 42.12 -> 51.39 | Helped |
Closest papers by meaning
- Towards Global Video Scene Segmentation with Context-Aware Transformer (AAAI 2023 · similarity 0.90)
- Video Scene Segmentation with Genre and Duration Signals (ICLR 2026 · similarity 0.89)
- Learning From Untrimmed Videos: Self-Supervised Video Representation Learning With Hierarchical Consistency (CVPR 2022 · similarity 0.88)
- Exploring Temporal Feature Correlation for Efficient and Stable Video Semantic Segmentation (AAAI 2024 · similarity 0.88)
- Pay Attention to Target: Relation-Aware Temporal Consistency for Domain Adaptive Video Semantic Segmentation (AAAI 2024 · similarity 0.87)
- Minimizing Labeled, Maximizing Unlabeled: An Image-Driven Approach for Video Instance Segmentation (CVPR 2025 · similarity 0.86)
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.