Research map / Vision-language and multimodal / Cross-modal learning
A Unified Multimodal De- and Re-Coupling Framework for RGB-D Motion Recognition
Benjia Zhou, Pichao Wang, Jun Wan, Yanyan Liang, Fan Wang
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 multi-stage recoupling (3 RCM modules instead of 1) (ablation) | UMDR · NTU RGB-D CS (RGB, 32 frames) | accuracy: 91.46 -> 92.24 | Improved |
| add multi-stage recoupling (3 RCM modules instead of 1) (ablation) | UMDR · NvGesture (RGB, 32 frames) | accuracy: 84.52 -> 86.25 | Improved |
| add multi-stage recoupling (3 RCM modules instead of 1) (ablation) | UMDR · NTU RGB-D CS (RGB, 16 frames) | accuracy: 90.81 -> 91.37 | Improved |
| add recoupling self-distillation loss (lambda_d 0.2) (ablation) | UMDR · NvGesture (RGB, 32 frames) | accuracy: 85.04 -> 86.25 | Improved |
| add recoupling self-distillation loss (lambda_d 0.2) (ablation) | UMDR · NTU RGB-D CS (RGB, 32 frames) | accuracy: 91.56 -> 92.24 | Improved |
| raise recoupling loss weight from 0.2 to 0.3 (ablation) | UMDR · NTU RGB-D CS (RGB, 32 frames) | accuracy: 92.24 -> 91.87 | Hurt |
| use batch-wise instead of pair-wise ShuffleMix+ (ablation) | UMDR · NTU RGB-D CS (RGB, 32 frames) | accuracy: 91.41 -> 92.24 | Improved |
| use discrete instead of continuous temporal mixing (ablation) | UMDR · IsoGD (RGB, 32 frames) | accuracy: 63.18 -> 63.68 | Improved |
| use continuous instead of discrete temporal mixing (ablation) | UMDR · NvGesture (RGB, 32 frames) | accuracy: 86.25 -> 86.88 | Improved |
| add MixUp alone (alpha_m 0.8) to baseline | UMDR · IsoGD (RGB) | accuracy: +7.63% | Improved |
| add MixUp alone to baseline on small-scale data (ablation) | UMDR · NvGesture | accuracy | No clear effect |
| add ShuffleMix to MixUp (ShuffleMix+) on small datasets (ablation) | UMDR · NvGesture | accuracy: +2.1% | Improved |
Closest papers by meaning
- Decoupling and Recoupling Spatiotemporal Representation for RGB-D-Based Motion Recognition (CVPR 2022 · similarity 0.90)
- Deep Multimodal Feature Analysis for Action Recognition in RGB+D Videos (TPAMI 2017 · similarity 0.88)
- MMNet: A Model-Based Multimodal Network for Human Action Recognition in RGB-D Videos (TPAMI 2022 · similarity 0.88)
- Decomposed Cross-Modal Distillation for RGB-Based Temporal Action Detection (CVPR 2023 · similarity 0.86)
- Don't Judge by the Look: Towards Motion Coherent Video Representation (ICLR 2024 · similarity 0.85)
- Keep the Balance: A Parameter-Efficient Symmetrical Framework for RGB+X Semantic Segmentation (CVPR 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.
