Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Personalizing Gesture Recognition Using Hierarchical Bayesian Neural Networks
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 a hierarchical Bayesian neural network instead of hierarchical Bayesian multinomial regression | Gesture recognition with known subject membership · MSRC-12, ChaLearn 2013, and NATOPS | F1-score | Helped |
| Use subject-specific hierarchical models instead of pooling all subjects into one Bayesian neural network | HBNN · MSRC-12, ChaLearn 2013, and NATOPS | F1-score | Helped |
| Use hierarchical multinomial regression instead of a pooled Bayesian neural network | Gesture recognition · MSRC-12 | F1-score | Helped |
| Use hierarchical multinomial regression instead of a pooled Bayesian neural network | Gesture recognition · ChaLearn 2013 and NATOPS | F1-score | Hurt |
| Use weighted Bayesian model averaging instead of naive uniform Bayesian model averaging (ablation) | HBNN with unknown subject membership · MSRC-12 and NATOPS | F1-score | Helped |
| Use weighted Bayesian model averaging instead of naive uniform Bayesian model averaging (ablation) | HBNN with unknown subject membership · ChaLearn 2013 | F1-score | No clear effect |
| Use known subject membership instead of inferring membership (ablation) | HBNN · MSRC-12 and NATOPS | F1-score | Helped |
| Use known subject membership instead of inferring membership (ablation) | HBNN · ChaLearn 2013 | F1-score | No clear effect |
| Select personalization examples with BALD instead of at random (ablation) | HBNN personalization · MSRC-12, when personalization instances per gesture exceed 1 | F1-score | Helped |
| Select personalization examples with BALD instead of at random (ablation) | HBNN personalization · NATOPS, when personalization instances per gesture exceed 3 | F1-score | Helped |
| Select personalization examples with BALD instead of at random (ablation) | HBNN personalization · ChaLearn 2013, when personalization instances per gesture exceed 4 | F1-score | Helped |
| Use BALD instead of random selection with very few personalization examples (ablation) | HBNN personalization · MSRC-12, NATOPS, and ChaLearn 2013 | F1-score | No clear effect |
Closest papers by meaning
- Online Detection and Classification of Dynamic Hand Gestures With Recurrent 3D Convolutional Neural Network (CVPR 2016 · similarity 0.87)
- SocialGesture: Delving into Multi-person Gesture Understanding (CVPR 2025 · similarity 0.83)
- Hand1000: Generating Realistic Hands from Text with Only 1,000 Images (AAAI 2025 · similarity 0.82)
- Expressive Forecasting of 3D Whole-Body Human Motions (AAAI 2024 · similarity 0.82)
- Distribution Matching for Multi-Task Learning of Classification Tasks: A Large-Scale Study on Faces & Beyond (AAAI 2024 · similarity 0.82)
- DynASyn: Multi-Subject Personalization Enabling Dynamic Action Synthesis (AAAI 2025 · similarity 0.81)
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.