Research map / Video / Action recognition
FineBio: A Fine-Grained Video Dataset of Biological Experiments with Hierarchical Annotation
Takuma Yagi, Misaki Ohashi, Yifei Huang, Ryosuke Furuta, Shungo Adachi, Toutai Mitsuyama, Yoichi Sato
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 set-classification head on top of separate heads (ablation) | ActionFormer · FineBio atomic operation detection | avg mAP atomic operation: 28.3 -> 34.1 | Improved |
| add separate entity heads on top of set classification (ablation) | ActionFormer · FineBio atomic operation detection | avg mAP atomic operation: 31.9 -> 34.1 | Improved |
| add set-classification head to separate heads (ablation) | ActionFormer · FineBio verb detection | avg mAP verb: 58.5 -> 60.0 | Improved |
| share weights across entities instead of separate models (ablation) | ActionFormer · FineBio atomic operation detection | avg mAP atomic operation: 28.1 -> 34.1 | Improved |
| add set-classification head to separate heads (ablation) | ActionFormer · FineBio manipulated object detection | avg mAP manipulated object: 41.9 -> 41.6 | No clear effect |
| use DINO instead of Deformable DETR | object detector · FineBio object detection | AP: 53.3 -> 56.1 | Improved |
| use DINO instead of Deformable DETR | object detector · FineBio object detection | AR manipulated: 55.9 -> 64.0 | Improved |
| use MS-TCN++ instead of ASFormer | step segmentation · FineBio step segmentation | F1@75: 67.0 -> 73.4 | Improved |
| raise tIoU threshold from F1@50 to F1@75 | MS-TCN++ · FineBio step segmentation | F1: 93.5 -> 73.4 | Hurt |
| require manipulated and affected objects in addition to hands | Hand Object Detector with DINO · FineBio manipulated/affected object detection (right hand) | AP50: 94.5 -> 22.2 | Hurt |
Closest papers by meaning
- ProBio: A Protocol-guided Multimodal Dataset for Molecular Biology Lab (NeurIPS 2023 · similarity 0.88)
- AVA: A Video Dataset of Spatio-Temporally Localized Atomic Visual Actions (CVPR 2018 · similarity 0.85)
- Weakly-Supervised Temporal Action Detection for Fine-Grained Videos with Hierarchical Atomic Actions (ECCV 2022 · similarity 0.85)
- FineGym: A Hierarchical Video Dataset for Fine-Grained Action Understanding (CVPR 2020 · similarity 0.84)
- FineParser: A Fine-grained Spatio-temporal Action Parser for Human-centric Action Quality Assessment (CVPR 2024 · similarity 0.84)
- FineSports: A Multi-person Hierarchical Sports Video Dataset for Fine-grained Action Understanding (CVPR 2024 · similarity 0.83)
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.
