Research map / Vision-language and multimodal / Cross-modal learning
Human Activity Recognition using RGB-DVS Cameras: A Multi-modal Heat Conduction Model and A Benchmark Dataset
Shiao Wang, Xiao Wang, Bo Jiang, Lin Zhu, Guoqi Li, Yaowei Wang, Yonghong Tian, Jin Tang
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 |
|---|---|---|---|
| replace Vision Mamba with multi-modal heat conduction model | Vision Mamba · HARDVS 2.0 | top-1 accuracy: 51.8 -> 53.2 | Improved |
| add modality-specific continuous frequency value embeddings (ablation) | vHeat · HARDVS 2.0 | top-1 accuracy: 52.3 -> 52.8 | Improved |
| add policy routing based fusion (ablation) | vHeat · HARDVS 2.0 | top-1 accuracy: 52.3 -> 52.7 | Improved |
| combine MSC and PRF (ablation) | vHeat · HARDVS 2.0 | top-1 accuracy: 52.3 -> 53.2 | Improved |
| add event streams to RGB frames (ablation) | RGB-only model · HARDVS 2.0 | top-1 accuracy: 49.2 -> 53.2 | Improved |
| add RGB frames to event streams (ablation) | event-only model · HARDVS 2.0 | top-1 accuracy: 51.2 -> 53.2 | Improved |
| event-only input versus RGB-only input (ablation) | single-modality model · HARDVS 2.0 | top-1 accuracy: 49.2 -> 51.2 | Improved |
| use adaptive routing selection instead of simple addition (ablation) | fusion · HARDVS 2.0 | top-1 accuracy: 52.3 -> 53.2 | Improved |
| use adaptive routing instead of random selection among three fusion methods (ablation) | fusion · HARDVS 2.0 | top-1 accuracy: 52.6 -> 53.2 | Improved |
| use concatenation instead of simple addition (ablation) | fusion · HARDVS 2.0 | top-1 accuracy: 52.3 -> 52.3 | No clear effect |
| raise input resolution from 224x224 to 256x256 (ablation) | multi-modal HCO · HARDVS 2.0 | top-1 accuracy | Hurt |
| use 8 input frames rather than 4 or 12 (ablation) | multi-modal HCO · HARDVS 2.0 | top-1 accuracy | Improved |
Closest papers by meaning
- HARDVS: Revisiting Human Activity Recognition with Dynamic Vision Sensors (AAAI 2024 · similarity 0.84)
- NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysis (CVPR 2016 · similarity 0.83)
- NTU RGB+D 120: A Large-Scale Benchmark for 3D Human Activity Understanding (TPAMI 2019 · similarity 0.82)
- Heatmap Pooling Network for Action Recognition From RGB Videos (TPAMI 2025 · similarity 0.82)
- Multi-Modal Hand-to-Mouth Gesture Recognition in Activity-Oriented RGB-Thermal Footage (Student Abstract) (AAAI 2025 · similarity 0.82)
- Decomposing and Fusing Intra- and Inter-Sensor Spatio-Temporal Signal for Multi-Sensor Wearable Human Activity Recognition (AAAI 2025 · similarity 0.80)
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.
