NSTL国家科技图书文献中心

1. Extended multi-stream temporal-attention module for skeleton-based human action recognition (HAR) NSTL国家科技图书文献中心

Mehmood F. | Guo X.... - 《Computers in human behavior》 - 2025,163(Feb.) - 1.1～1.14 - 共14页 - 被引量：1

摘要： action recognition (HAR) technique. GCNs enable the | significance and variance in HAR. For this purpose, we | demonstrating the superiority of our model for the task of HAR | © 2024 Elsevier LtdGraph convolutional | networks (GCNs) are an effective skeleton-based human

关键词： Adaptive graph | CNN | HAR | Skeleton-based | Temporal module

2. Equimetrics - Applying HAR Principles to Equestrian Activities NSTL国家科技图书文献中心

Jonas Pohler | Kristof Van Laerhove... - 《Sensor-Based Activity Recognition and Artificial Intelligence》 - International Workshop on Sensor-Based Activity Recognition and Artificial Intelligence - 2025, - 69～78 - 共10页

摘要： capture system. The primary objective is to apply HAR | This paper presents the Equimetrics data | principles to enhance the understanding and optimization of | equestrian performance. By integrating data from | strategically placed sensors on the rider's body and the

关键词： Activity recognition | Equestrian activities | Transformer

3. ASK-HAR: Attention-Based Multi-Core Selective Kernel Convolution Network for Human Activity Recognition NSTL国家科技图书文献中心

Yu, Xugao | Al-qaness, Mohammed ... - 《Measurement》 - 2025,242PC - 115981～ - 共11页

摘要：Human Activity Recognition (HAR) is an | a prevalent sensor modality for HAR, offering | model, ASK-HAR* (Attention-based Multi Core Selective* | Kernel Convolution Network for HAR* ), which enhances* | HAR by performing kernel selection between multiple

关键词： Deep learning | Human activity recognition | Sensors | Selective kernel convolution network | Attention mechanisms

4. ALS-HAR: Harnessing Wearable Ambient Light Sensors to Enhance IMU-Based Human Activity Recognition NSTL国家科技图书文献中心

Lala Shakti Swarup R... | Daniel Geissler ... - 《Pattern Recognition,Part XXIX》 - International Conference on Pattern Recognition - 2025, - 133～147 - 共15页

摘要： human activity recognition (HAR), primarily through | , we developed ALS-HAR, a robust wearable light-based | motion activity classifier. Although ALS-HAR* achieves* | , cross-modal information can still improve other HAR | of ALS integration in advancing sensor-based HAR

关键词： Human activity recognition | Ambient light sensor | Sensor fusion | Contrastive learning | IMU sensing

5. S~2C-HAR: A Semi-Supervised Human Activity Recognition Framework Based on Contrastive Learning NSTL国家科技图书文献中心

Xue Li | Mingxing Liu ... - 《Concurrency and computation: practice and experience》 - 2025,37(6/8) - e70027.1～e70027.18 - 共18页

摘要：Human activity recognition (HAR) has emerged | , existing HAR* methods predominantly rely on deep* | paper, we propose a novel semi-supervised HAR | framework based on contrastive learning named S2C-HAR | applied. First, a contrastive learning model for HAR

关键词： contrastive learning | human activity recognition | semi-supervised learning

6. Cross-Domain HAR: Few-Shot Transfer Learning for Human Activity Recognition NSTL国家科技图书文献中心

MEGHA THUKRAL | HARISH HARESAMUDRAM ... - 《ACM transactions on intelligent systems and technology》 - 2025,16(1) - 22.1～22.35 - 共35页

摘要： applications of sensor-based human activity recognition (HAR | effective HAR* systems and less than ideal recognition* | HAR systems. Yet they can fail when the differences | economic use of publicly available labeled HAR* datasets* | locations and type of activities. Cross- Domain HAR

关键词： human activity recognition | representation learning | self-training | selfsupervised learning | semi-supervised learning

7. Refined Human Activity Recognition Algorithm Using Oversampling and Combinatorial Loss in 1DCLA-Net NSTL国家科技图书文献中心

Zeting Xu | Quan Liu ... - 《Intelligent Robotics and Applications,Part V》 - International Conference on Intelligent Robotics and Applications - 2025, - 204～216 - 共13页

摘要： recognition (HAR) has become mainstream, especially playing | , existing HAR* algorithms based on deep learning often* | % and 97.87% on UCI HAR* and WISDM, outperforming* | With the rapid development of the artificial | intelligence field today, sensor-based human activity

关键词： Deep learning | HAR | 1DCNN | LSTM | Attention mechanism

8. Parallel Attention Based Network for Human Activity Recognition Using Wearable Devices NSTL国家科技图书文献中心

Chenyang Xu | Feiyi Fan ... - 《Pattern Recognition,Part XIII》 - International Conference on Pattern Recognition - 2025, - 450～465 - 共16页

摘要： activity recognition (HAR) tasks using deep learning | paper, we present the parallel attention-based HAR* (PA* | -HAR) method. Our method employs multiple small-scale | datasets show that our proposed PA-HAR* model outperforms* | state-of-the-art approaches in HAR* tasks. In addition*

关键词： Human activity recognition (HAR) | Wearable device | Parallel attention (PA)

9. A Multiple Wi-Fi Sensors Assisted Human Activity Recognition Scheme for Smart Home NSTL国家科技图书文献中心

Ding, Jianyang | Wang, Yong ... - 《IEEE sensors journal》 - 2025,25(3) - 4958～4968 - 共11页

摘要： activity recognition (HAR) relying on Wi-Fi signals has | HAR solutions fail to capture the full scope of the | separable convolution neural network (DSCNN)-based HAR* by* | calibration. We design and test an HAR* prototype on* | HAR compared with existing ones.

关键词： Channel state information (CSI) | human activity recognition (HAR) | smart healthcare | smart home | Wi-Fi sensors

10. Can climate factors improve the forecasting of electricity price volatility? Evidence from Australia NSTL国家科技图书文献中心

Kun Guo | Yu Liu ... - 《Energy》 - 2025,315(Jan..15) - 134332.1～134332.18 - 共18页

摘要： autoregressive realized volatility (HAR-RV) models. The results | market. Compared to the widely studied HAR-RV model and | The increasing share of renewable energy | generation and the frequent occurrence of extreme climate | events have made the impact of climate factors on

摘要： action recognition (HAR) technique. GCNs enable the | significance and variance in HAR. For this purpose, we | demonstrating the superiority of our model for the task of HAR | © 2024 Elsevier LtdGraph convolutional | networks (GCNs) are an effective skeleton-based human

关键词： Adaptive graph | CNN | HAR | Skeleton-based | Temporal module

摘要： capture system. The primary objective is to apply HAR | This paper presents the Equimetrics data | principles to enhance the understanding and optimization of | equestrian performance. By integrating data from | strategically placed sensors on the rider's body and the

关键词： Activity recognition | Equestrian activities | Transformer

摘要：Human Activity Recognition (HAR) is an | a prevalent sensor modality for HAR, offering | model, ASK-HAR* (Attention-based Multi Core Selective* | Kernel Convolution Network for HAR* ), which enhances* | HAR by performing kernel selection between multiple

关键词： Deep learning | Human activity recognition | Sensors | Selective kernel convolution network | Attention mechanisms

摘要： human activity recognition (HAR), primarily through | , we developed ALS-HAR, a robust wearable light-based | motion activity classifier. Although ALS-HAR* achieves* | , cross-modal information can still improve other HAR | of ALS integration in advancing sensor-based HAR

关键词： Human activity recognition | Ambient light sensor | Sensor fusion | Contrastive learning | IMU sensing

摘要：Human activity recognition (HAR) has emerged | , existing HAR* methods predominantly rely on deep* | paper, we propose a novel semi-supervised HAR | framework based on contrastive learning named S2C-HAR | applied. First, a contrastive learning model for HAR

关键词： contrastive learning | human activity recognition | semi-supervised learning

摘要： applications of sensor-based human activity recognition (HAR | effective HAR* systems and less than ideal recognition* | HAR systems. Yet they can fail when the differences | economic use of publicly available labeled HAR* datasets* | locations and type of activities. Cross- Domain HAR

关键词： human activity recognition | representation learning | self-training | selfsupervised learning | semi-supervised learning

摘要： recognition (HAR) has become mainstream, especially playing | , existing HAR* algorithms based on deep learning often* | % and 97.87% on UCI HAR* and WISDM, outperforming* | With the rapid development of the artificial | intelligence field today, sensor-based human activity

关键词： Deep learning | HAR | 1DCNN | LSTM | Attention mechanism

摘要： activity recognition (HAR) tasks using deep learning | paper, we present the parallel attention-based HAR* (PA* | -HAR) method. Our method employs multiple small-scale | datasets show that our proposed PA-HAR* model outperforms* | state-of-the-art approaches in HAR* tasks. In addition*

关键词： Human activity recognition (HAR) | Wearable device | Parallel attention (PA)

摘要： activity recognition (HAR) relying on Wi-Fi signals has | HAR solutions fail to capture the full scope of the | separable convolution neural network (DSCNN)-based HAR* by* | calibration. We design and test an HAR* prototype on* | HAR compared with existing ones.

关键词： Channel state information (CSI) | human activity recognition (HAR) | smart healthcare | smart home | Wi-Fi sensors

摘要： autoregressive realized volatility (HAR-RV) models. The results | market. Compared to the widely studied HAR-RV model and | The increasing share of renewable energy | generation and the frequent occurrence of extreme climate | events have made the impact of climate factors on

关键词： Electricity price volatility | Climate factor | HAR-RV model | Random forest

4008-161-200 800-990-8900

国家科技图书文献中心

摘要： action recognition (HAR) technique. GCNs enable the | significance and variance in HAR. For this purpose, we | demonstrating the superiority of our model for the task of HAR | © 2024 Elsevier LtdGraph convolutional | networks (GCNs) are an effective skeleton-based human

关键词： Adaptive graph | CNN | HAR | Skeleton-based | Temporal module

摘要： capture system. The primary objective is to apply HAR | This paper presents the Equimetrics data | principles to enhance the understanding and optimization of | equestrian performance. By integrating data from | strategically placed sensors on the rider's body and the

关键词： Activity recognition | Equestrian activities | Transformer

摘要：Human Activity Recognition (HAR) is an | a prevalent sensor modality for HAR, offering | model, ASK-HAR (Attention-based Multi Core Selective | Kernel Convolution Network for HAR ), which enhances | HAR by performing kernel selection between multiple

关键词： Deep learning | Human activity recognition | Sensors | Selective kernel convolution network | Attention mechanisms

摘要： human activity recognition (HAR), primarily through | , we developed ALS-HAR, a robust wearable light-based | motion activity classifier. Although ALS-HAR achieves | , cross-modal information can still improve other HAR | of ALS integration in advancing sensor-based HAR

关键词： Human activity recognition | Ambient light sensor | Sensor fusion | Contrastive learning | IMU sensing

摘要：Human activity recognition (HAR) has emerged | , existing HAR methods predominantly rely on deep | paper, we propose a novel semi-supervised HAR | framework based on contrastive learning named S2C-HAR | applied. First, a contrastive learning model for HAR

关键词： contrastive learning | human activity recognition | semi-supervised learning

摘要： applications of sensor-based human activity recognition (HAR | effective HAR systems and less than ideal recognition | HAR systems. Yet they can fail when the differences | economic use of publicly available labeled HAR datasets | locations and type of activities. Cross- Domain HAR

关键词： human activity recognition | representation learning | self-training | selfsupervised learning | semi-supervised learning

摘要： recognition (HAR) has become mainstream, especially playing | , existing HAR algorithms based on deep learning often | % and 97.87% on UCI HAR and WISDM, outperforming | With the rapid development of the artificial | intelligence field today, sensor-based human activity

关键词： Deep learning | HAR | 1DCNN | LSTM | Attention mechanism

摘要： activity recognition (HAR) tasks using deep learning | paper, we present the parallel attention-based HAR (PA | -HAR) method. Our method employs multiple small-scale | datasets show that our proposed PA-HAR model outperforms | state-of-the-art approaches in HAR tasks. In addition

关键词： Human activity recognition (HAR) | Wearable device | Parallel attention (PA)

摘要： activity recognition (HAR) relying on Wi-Fi signals has | HAR solutions fail to capture the full scope of the | separable convolution neural network (DSCNN)-based HAR by | calibration. We design and test an HAR prototype on | HAR compared with existing ones.

关键词： Channel state information (CSI) | human activity recognition (HAR) | smart healthcare | smart home | Wi-Fi sensors

摘要： autoregressive realized volatility (HAR-RV) models. The results | market. Compared to the widely studied HAR-RV model and | The increasing share of renewable energy | generation and the frequent occurrence of extreme climate | events have made the impact of climate factors on

关键词： Electricity price volatility | Climate factor | HAR-RV model | Random forest

4008-161-200 800-990-8900

国家科技图书文献中心

摘要：Human Activity Recognition (HAR) is an | a prevalent sensor modality for HAR, offering | model, ASK-HAR* (Attention-based Multi Core Selective* | Kernel Convolution Network for HAR* ), which enhances* | HAR by performing kernel selection between multiple

摘要： human activity recognition (HAR), primarily through | , we developed ALS-HAR, a robust wearable light-based | motion activity classifier. Although ALS-HAR* achieves* | , cross-modal information can still improve other HAR | of ALS integration in advancing sensor-based HAR

摘要：Human activity recognition (HAR) has emerged | , existing HAR* methods predominantly rely on deep* | paper, we propose a novel semi-supervised HAR | framework based on contrastive learning named S2C-HAR | applied. First, a contrastive learning model for HAR

摘要： applications of sensor-based human activity recognition (HAR | effective HAR* systems and less than ideal recognition* | HAR systems. Yet they can fail when the differences | economic use of publicly available labeled HAR* datasets* | locations and type of activities. Cross- Domain HAR

摘要： recognition (HAR) has become mainstream, especially playing | , existing HAR* algorithms based on deep learning often* | % and 97.87% on UCI HAR* and WISDM, outperforming* | With the rapid development of the artificial | intelligence field today, sensor-based human activity

摘要： activity recognition (HAR) tasks using deep learning | paper, we present the parallel attention-based HAR* (PA* | -HAR) method. Our method employs multiple small-scale | datasets show that our proposed PA-HAR* model outperforms* | state-of-the-art approaches in HAR* tasks. In addition*

摘要： activity recognition (HAR) relying on Wi-Fi signals has | HAR solutions fail to capture the full scope of the | separable convolution neural network (DSCNN)-based HAR* by* | calibration. We design and test an HAR* prototype on* | HAR compared with existing ones.