Research map / Video / Action recognition
Soft Video Parsing by Label Distribution Learning
Xin Geng, Miaogen Ling
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 |
|---|---|---|---|
| replace hard single-label parsing (SSRG) with soft label distribution (SP) | SSRG · THUMOS'14 average | AP: 10.8 -> 18.3 | Improved |
| replace SSRG with SP | SSRG · THUMOS'14 average | frame labeling accuracy: 52.8 -> 66.0 | Improved |
| replace SSRG with SP | SSRG · MSR-II average | AP: 9.0 -> 23.9 | Improved |
| replace SSRG with SP | SSRG · MSR-II average | frame labeling accuracy: 66.7 -> 77.9 | Improved |
| add sub-action decomposition (SSRG/SP) versus sliding-window RNMS | RNMS · THUMOS'14 average | AP: 7.8 -> 10.8 | Improved |
| replace SSRG with SP on billiards | SSRG · THUMOS'14 billiards | AP: 3.7 -> 3.1 | Hurt |
| replace SSRG with SP on tennis-swing | SSRG · THUMOS'14 tennis-swing | frame labeling accuracy: 66.9 -> 58.4 | Hurt |
| replace RNMS with SP on long-jump | RNMS · THUMOS'14 long-jump | AP: 35.3 -> 26.7 | Hurt |
| tune ambiguity level lambda per action (OAL) instead of fixed 1.0 (ablation) | SP · THUMOS'14 average | AP: 18.3 -> 22.1 | Improved |
| tune ambiguity level lambda per action (OAL) (ablation) | SP · THUMOS'14 average | frame labeling accuracy: 66.0 -> 68.1 | Improved |
Closest papers by meaning
- Advancing Weakly-Supervised Audio-Visual Video Parsing via Segment-Wise Pseudo Labeling (IJCV 2024 · similarity 0.86)
- Revisit Weakly-Supervised Audio-Visual Video Parsing from the Language Perspective (NeurIPS 2023 · similarity 0.86)
- Unsupervised Semantic Scene Labeling for Streaming Data (CVPR 2017 · similarity 0.85)
- Action Sets: Weakly Supervised Action Segmentation Without Ordering Constraints (CVPR 2018 · similarity 0.85)
- Learning from Weakly-Labeled Web Videos via Exploring Sub-concepts (AAAI 2022 · similarity 0.85)
- Weakly-Supervised Audio-Visual Video Parsing with Prototype-based Pseudo-Labeling (CVPR 2024 · 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.
