Research map / Video / Action recognition
Prediction of Manipulation Actions
Cornelia M. Fermuller, Fang Wang, Yezhou Yang, Konstantinos Zampogiannis, Yi Zhang, Francisco Barranco, Michael Pfeiffer
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 |
|---|---|---|---|
| use VGG16 CNN features instead of HOG features as LSTM input (ablation) | LSTM action classifier · MAD action classification | accuracy: 59.2% -> 68.3% | Improved |
| use LSTM with VGG16 features instead of HMM | action classifier · MAD action classification | accuracy: 53.7% -> 68.3% | Improved |
| use LSTM with VGG16 features instead of SVM | action classifier · MAD action classification | accuracy: 48.1% -> 68.3% | Improved |
| add predicted finger forces to vision features | action prediction LSTM · MAD action prediction (average) | accuracy: 68.3% -> 70.5% | Improved |
| add predicted finger forces to vision features | action prediction LSTM · MAD stone | accuracy: 61.4% -> 75.1% | Improved |
| add predicted finger forces to vision features | action prediction LSTM · MAD cup | accuracy: 82.4% -> 88.2% | Improved |
| add predicted finger forces to vision features | action prediction LSTM · MAD sponge | accuracy: 61.6% -> 59.1% | Hurt |
| add predicted finger forces to vision features | action prediction LSTM · MAD spoon | accuracy: 62.6% -> 57.5% | Hurt |
| add predicted finger forces to vision features | action prediction LSTM · MAD knife | accuracy: 73.3% -> 72.7% | No clear effect |
| let human observers see more frames after contact (10 to 25 frames) | human action prediction · MAD sponge human study | accuracy: about 75% -> above 95% for four of five actions | Improved |
| give human observers feedback on correct action | human action prediction · MAD human study, spoon at 10 frames after contact | accuracy | Improved |
Closest papers by meaning
- Forecasting Action Through Contact Representations From First Person Video (TPAMI 2021 · similarity 0.89)
- World Models for Learning Dexterous Hand-Object Interactions from Human Videos (ECCV 2026 · similarity 0.88)
- Dexterous Manipulation Transfer via Progressive Kinematic-Dynamic Alignment (AAAI 2026 · similarity 0.87)
- ManipTrans: Efficient Dexterous Bimanual Manipulation Transfer via Residual Learning (CVPR 2025 · similarity 0.85)
- Forecasting Bimanual Object Manipulation Sequences from Unimanual Observations (AAAI 2024 · similarity 0.85)
- Robots Pre-train Robots: Manipulation-Centric Robotic Representation from Large-Scale Robot Datasets (ICLR 2025 · 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.
