Research map / Video / Video segmentation and tracking
Complex Event Detection by Identifying Reliable Shots From Untrimmed Videos
Hehe Fan, Xiaojun Chang, De Cheng, Yi Yang, Dong Xu, Alexander G. Hauptmann
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 |
|---|---|---|---|
| select reliable shots with MIL-SRI instead of treating all shots as instances | SVM · MEDTest-14 100Ex | mAP: 22.2 -> 38.6 | Improved |
| use MIL-SRI instead of SMIL-TopK | MIL-SRI · MEDTest-14 100Ex | mAP: 36.9 -> 38.6 | Improved |
| use MIL-SRI instead of SMIL-TopK | MIL-SRI · MEDTest-13 100Ex | mAP: 40.5 -> 43.1 | Improved |
| pool frame features with LSTM instead of sparse MIL | RNN-LSTM · MEDTest-14 100Ex | mAP: 19.8 -> 31.1 | Improved |
| add l2 diversity term (gamma > 0) (ablation) | MIL-SRI · MEDTest-14 (rock climbing E027, 100Ex) | AP: 30.76% -> 32.64% | Improved |
| add IDT motion feature fusion | MIL-SRI · MEDTest-14 10Ex | mAP: 28.4 -> 29.6 | Improved |
| add IDT motion feature fusion | MIL-SRI · MEDTest-13 100Ex | mAP: 43.1 -> 49.7 | Improved |
| use MIL-SRI instead of NI-SVM | MIL-SRI · CCV | mAP: 78.3 -> 77.9 | Hurt |
| use MIL-SRI instead of CNN + VLAD | MIL-SRI · MEDTest-14 100Ex | mAP: 35.7 -> 38.6 | Improved |
Closest papers by meaning
- Weakly-Supervised Video Anomaly Detection With Robust Temporal Feature Magnitude Learning (ICCV 2021 · similarity 0.86)
- Prompt-Enhanced Multiple Instance Learning for Weakly Supervised Video Anomaly Detection (CVPR 2024 · similarity 0.86)
- Semantic Pooling for Complex Event Analysis in Untrimmed Videos (TPAMI 2016 · similarity 0.86)
- They Are Not Equally Reliable: Semantic Event Search Using Differentiated Concept Classifiers (CVPR 2016 · similarity 0.86)
- Learning Latent Super-Events to Detect Multiple Activities in Videos (CVPR 2018 · similarity 0.85)
- Learning Event Completeness for Weakly Supervised Video Anomaly Detection (ICML 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.
