Confidence-driven Gradient Modulation for Multimodal Human Activity Recognition: A Dynamic Contrastive Dual-Path Learning Approach
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866914004925939712 |
|---|---|
| author | Ji, Panpan Song, Junni Lu, Yifan Xiao, Hang Liu, Hanyu Li, Chao |
| author_facet | Ji, Panpan Song, Junni Lu, Yifan Xiao, Hang Liu, Hanyu Li, Chao |
| contents | Sensor-based Human Activity Recognition (HAR) is a core technology that enables intelligent systems to perceive and interact with their environment. However, multimodal HAR systems still encounter key challenges, such as difficulties in cross-modal feature alignment and imbalanced modality contributions. To address these issues, we propose a novel framework called the Dynamic Contrastive Dual-Path Network (DCDP-HAR). The framework comprises three key components. First, a dual-path feature extraction architecture is employed, where ResNet and DenseNet branches collaboratively process multimodal sensor data. Second, a multi-stage contrastive learning mechanism is introduced to achieve progressive alignment from local perception to semantic abstraction. Third, we present a confidence-driven gradient modulation strategy that dynamically monitors and adjusts the learning intensity of each modality branch during backpropagation, effectively alleviating modality competition. In addition, a momentum-based gradient accumulation strategy is adopted to enhance training stability. We conduct ablation studies to validate the effectiveness of each component and perform extensive comparative experiments on four public benchmark datasets. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_02826 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Confidence-driven Gradient Modulation for Multimodal Human Activity Recognition: A Dynamic Contrastive Dual-Path Learning Approach Ji, Panpan Song, Junni Lu, Yifan Xiao, Hang Liu, Hanyu Li, Chao Computer Vision and Pattern Recognition Sensor-based Human Activity Recognition (HAR) is a core technology that enables intelligent systems to perceive and interact with their environment. However, multimodal HAR systems still encounter key challenges, such as difficulties in cross-modal feature alignment and imbalanced modality contributions. To address these issues, we propose a novel framework called the Dynamic Contrastive Dual-Path Network (DCDP-HAR). The framework comprises three key components. First, a dual-path feature extraction architecture is employed, where ResNet and DenseNet branches collaboratively process multimodal sensor data. Second, a multi-stage contrastive learning mechanism is introduced to achieve progressive alignment from local perception to semantic abstraction. Third, we present a confidence-driven gradient modulation strategy that dynamically monitors and adjusts the learning intensity of each modality branch during backpropagation, effectively alleviating modality competition. In addition, a momentum-based gradient accumulation strategy is adopted to enhance training stability. We conduct ablation studies to validate the effectiveness of each component and perform extensive comparative experiments on four public benchmark datasets. |
| title | Confidence-driven Gradient Modulation for Multimodal Human Activity Recognition: A Dynamic Contrastive Dual-Path Learning Approach |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2507.02826 |