Research map / Video / Action recognition
Learning to Discriminate Information for Online Action Detection: Analysis and Application
Sumin Lee, Hyunjun Eun, Jinyoung Moon, Seokeon Choi, Yoonhyung Kim, Chanho Jung, Changick 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 current information input to GRU (ablation) | GRU · TVSeries | mcAP: 81.3 -> 83.4 | Improved |
| add early embedding module to GRU+CI (full IDN) (ablation) | Baseline+CI · TVSeries | mcAP: 83.4 -> 84.7 | Improved |
| add current information input to GRU (ablation) | GRU · THUMOS-14 | mAP: 46.7 -> 48.6 | Improved |
| add early embedding module to GRU+CI (full IDN) (ablation) | Baseline+CI · THUMOS-14 | mAP: 48.6 -> 50.0 | Improved |
| add both classification loss L_ee and contrastive loss L_ct to early embedding (ablation) | IDN · THUMOS-14 | mAP: 48.6 -> 50.0 | Improved |
| add both L_ee and L_ct to early embedding (ablation) | IDN · TVSeries | mcAP: 83.4 -> 84.7 | Improved |
| use GRU instead of simple RNN as recurrent unit (ablation) | online detection baseline · TVSeries | mcAP: 79.9 -> 81.3 | Improved |
| replace IDU with IIU for anticipation (ablation) | IIN · TVSeries 2.0s | mcAP: 70.3 -> 74.3 | Improved |
| replace GRU with IIU in anticipation network (ablation) | IIN · THUMOS-14 2.0s | mAP: 40.0 -> 47.2 | Improved |
| replace GRU with IIU in anticipation network (ablation) | IIN · TVSeries 1.0s | mcAP: 73.9 -> 76.7 | Improved |
| add h_{t-1} combination stage to IIU (ablation) | IIU · THUMOS-14 2.0s | mAP: 44.7 -> 47.2 | Improved |
| add weighted modulation stage to IIU concatenation (ablation) | IIU · THUMOS-14 1.0s | mAP: 48.8 -> 51.2 | Improved |
Closest papers by meaning
- Learning to Discriminate Information for Online Action Detection (CVPR 2020 · similarity 0.98)
- End-To-End Learning of Action Detection From Frame Glimpses in Videos (CVPR 2016 · similarity 0.88)
- Action Search: Spotting Actions in Videos and Its Application to Temporal Action Localization (ECCV 2018 · similarity 0.87)
- Temporal Recurrent Networks for Online Action Detection (ICCV 2019 · similarity 0.87)
- Action Detail Matters: Refining Video Recognition with Local Action Queries (CVPR 2025 · similarity 0.86)
- Text-Infused Attention and Foreground-Aware Modeling for Zero-Shot Temporal Action Detection (NeurIPS 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.
