675 Hits in 5.9 sec

Improving Deep Learning for HAR with shallow LSTMs [article]

Marius Bock, Alexander Hoelzemann, Michael Moeller, Kristof Van Laerhoven
2021 arXiv   pre-print
Recent studies in Human Activity Recognition (HAR) have shown that Deep Learning methods are able to outperform classical Machine Learning algorithms.  ...  One popular Deep Learning architecture in HAR is the DeepConvLSTM. In this paper we propose to alter the DeepConvLSTM architecture to employ a 1-layered instead of a 2-layered LSTM.  ...  of Deep Learning in HAR (e.g., [13, 26] ).  ... 
arXiv:2108.00702v2 fatcat:mko4bn7eezeunjesjjwltxuph4

Deep Architectures for Human Activity Recognition using Sensors

Zartasha Baloch, Faisal Karim Shaikh, Mukhtiar Ali Unar
2019 3C Tecnología  
Deep learning (DL) is becoming popular among HAR researchers due to its outstanding performance over conventional ML techniques.  ...  The use of sensors in the field of HAR opens new avenues for machine learning (ML) researchers to accurately recognize human activities.  ...  ACKNOWLEDGEMENTS This work has been performed under IICT, Mehran University of Engineering and Technology, Jamshoro and funded by ICT Endowment Fund for sustainable development.  ... 
doi:10.17993/3ctecno.2019.specialissue2.14-35 fatcat:cinqa2t2uvd6bklp3vor6ljylm

Communication-Efficient Federated Deep Learning with Asynchronous Model Update and Temporally Weighted Aggregation [article]

Yang Chen, Xiaoyan Sun, Yaochu Jin
2019 arXiv   pre-print
In the asynchronous learning strategy, different layers of the deep neural networks are categorized into shallow and deeps layers and the parameters of the deep layers are updated less frequently than  ...  The proposed algorithm is empirically on two datasets with different deep neural networks.  ...  ACKNOWLEDGMENT This work is supported by the National Natural Science Foundation of China with Grant No.61473298 and 61876184.  ... 
arXiv:1903.07424v1 fatcat:yu2hcwe64jgylef57q4etfhnym

A Hybrid CNN–LSTM Network for the Classification of Human Activities Based on Micro-Doppler Radar

JianPing Zhu, HaiQuan Chen, Wenbin Ye
2020 IEEE Access  
Fig. 1 shows the overall network architecture for HAR tasks, which consists of an STFT for data preprocessing, 1D-CNNs for local feature learning, an LSTM layer for global temporal information extraction  ...  In recent years, HAR has made remarkable advance by applying deep learning (DL) algorithms [12] - [18] to micro-Doppler signatures.  ...  Since 2015, he has been with the College of Electrical Science and Technology, Shenzhen University, where he is currently an Associate Professor.  ... 
doi:10.1109/access.2020.2971064 fatcat:xevslcdiyjf2tfjmrbf3yhxxii

Human Activity Recognition using Deep and Machine Learning Algorithms

Four deep learning approaches and thirteen different machine learning classifiers such as Multilayer Perceptron, Random Forest, Support Vector Machine, Decision Tree Classifier, AdaBoost Classifier, Gradient  ...  We have investigated all these classifiers to identify a best suitable classifier for this dataset.  ...  The performance difference between LSTM and Bi-LSTM on HAR was significantly minimal. The performance metrics of deep learning classifiers are shown in Fig. 2 for HAR. Fig. 2.  ... 
doi:10.35940/ijitee.c8835.029420 fatcat:flyl2djbcvh5rdra3pidpjb6nm

Basic Activity Recognition from Wearable Sensors Using a Lightweight Deep Neural Network

Zakaria Benhaili, Youness Abouqora, Youssef Balouki, Lahcen Moumoun
2022 Journal of ICT Standardization  
In this study, we propose a deep learning model that can work with raw data without any pre-processing. Several human activities can be recognized by our stacked LSTM network.  ...  Recently, the use of deep learning techniques allowed the extraction of features from sensor's readings automatically, in a hierarchical way through non-linear transformations.  ...  In [7] Agarwal et al. proposed a LSTM-CNN Architecture for Human Activity Recognition learning model for HAR.  ... 
doi:10.13052/jicts2245-800x.1028 dblp:journals/jicts/BenhailiABM22 fatcat:nn2gvwojbjcdxojb37ualy7qjq

A Comparative Analysis of Hybrid Deep Learning Models for Human Activity Recognition

Saedeh Abbaspour, Faranak Fotouhi, Ali Sedaghatbaf, Hossein Fotouhi, Maryam Vahabi, Maria Linden
2020 Sensors  
Deep learning (DL) is a branch of ML based on complex Artificial Neural Networks (ANNs) that has demonstrated a high level of accuracy and performance in HAR.  ...  We analyze four hybrid models that integrate CNNs with four powerful RNNs, i.e., LSTMs, BiLSTMs, GRUs and BiGRUs.  ...  The performance of LSTMs for real-time HAR is analyzed and compared with some other DL/ML models. LSTMs [45] The HAR performance of CNNs and that of CNN-LSTMs are compared.  ... 
doi:10.3390/s20195707 pmid:33036479 pmcid:PMC7582332 fatcat:otstrbwu4jfmbcp7u2h24xc54i

machine and deep learning approaches for human activity recognition

maha alhumayani, Mahmoud Monir, rasha ismail
2021 International Journal of Intelligent Computing and Information Sciences  
In this paper, a survey about the machine learning and deep learning methodologies in HAR is provided with information about the data, filtering methods, feature extraction methods, classification, and  ...  Then, the results obtained from the survey are discussed to explore the most efficient methods in both machine and deep learning for the recognition of HAR.  ...  Deep learning will improve the area of HAR by absorbing a huge amount of actions and activities for recognition.  ... 
doi:10.21608/ijicis.2021.82008.1106 fatcat:lcfqdcvtdnhfdfx3jvlvzq6kum

Joint Learning of Temporal Models to Handle Imbalanced Data for Human Activity Recognition

Rebeen Ali Hamad, Longzhi Yang, Wai Lok Woo, Bo Wei
2020 Applied Sciences  
Therefore, we aim to realise an activity recognition system using multi-modal sensors to address the issue of class imbalance in deep learning and improve recognition accuracy.  ...  This paper proposes a joint diverse temporal learning framework using Long Short Term Memory and one-dimensional Convolutional Neural Network models to improve human activity recognition, especially for  ...  with a large dataset for deep learning study.  ... 
doi:10.3390/app10155293 fatcat:b2eeu6pyy5e7vc4guc72mjch5y

Enhanced Hand-Oriented Activity Recognition Based on Smartwatch Sensor Data Using LSTMs

Sakorn Mekruksavanich, Anuchit Jitpattanakul, Phichai Youplao, Preecha Yupapin
2020 Symmetry  
This work proposes a hybrid deep learning model called CNN-LSTM that employed Long Short-Term Memory (LSTM) networks for activity recognition with the Convolution Neural Network (CNN).  ...  The results show that the proposed CNN-LSTM can support an improvement of the performance of activity recognition.  ...  Acknowledgments: The authors thank the SigOpt team for the provided optimization services. Conflicts of Interest: The authors declare no conflict of interest.  ... 
doi:10.3390/sym12091570 fatcat:4chkzpqh3jbzlfttiqcwzamhja

Deep Learning in Human Activity Recognition with Wearable Sensors: A Review on Advances

Shibo Zhang, Yaxuan Li, Shen Zhang, Farzad Shahabi, Stephen Xia, Yu Deng, Nabil Alshurafa
2022 Sensors  
We also present cutting-edge frontiers and future directions for deep learning-based HAR.  ...  Recently, deep learning has greatly pushed the boundaries of HAR on mobile and wearable devices.  ...  Acknowledgments: Special thanks to Haik Kalamtarian and Krystina Neuman for their valuable feedback. Conflicts of Interest: The authors declare no conflict of interest.  ... 
doi:10.3390/s22041476 pmid:35214377 pmcid:PMC8879042 fatcat:vp6jssypezbd5cnyzn4g35eqrm

Margin-Based Deep Learning Networks for Human Activity Recognition

Tianqi Lv, Xiaojuan Wang, Lei Jin, Yabo Xiao, Mei Song
2020 Sensors  
More recently, with significant progress in the development of deep learning networks for classification tasks, many researchers have made use of such models to recognise human activities in a sensor-based  ...  To deal with this problem, we introduce a margin mechanism to enhance the discriminative power of deep learning networks.  ...  In contrast to machine learning methods with shallow statistical features, deep learning methods can extract deep features automatically and have achieved superior performance in HAR.  ... 
doi:10.3390/s20071871 pmid:32230986 pmcid:PMC7181274 fatcat:4h5c667y45drvjq2w67cf42c2u

A Novel Multichannel Dilated Convolution Neural Network for Human Activity Recognition

Yingjie Lin, Jianning Wu
2020 Mathematical Problems in Engineering  
A novel multichannel dilated convolution neural network for improving the accuracy of human activity recognition is proposed.  ...  These results demonstrate that our model is an efficient real-time HAR model, which can gain the representative features from sensor signals at low computation and is hopeful for the effective tool in  ...  Related Work In recent years, various deep learning methods have been proposed for sensor-based HAR.  ... 
doi:10.1155/2020/5426532 fatcat:ekfnqzgvszgytj34ftarq3c2om

A Review of Deep Learning-based Human Activity Recognition on Benchmark Video Datasets

Vijeta Sharma, Manjari Gupta, Anil Kumar Pandey, Deepti Mishra, Ajai Kumar
2022 Applied Artificial Intelligence  
A short comparison is also made with the handcrafted feature-based approach and its fusion with deep learning to show the evolution of HAR methods.  ...  The objective of this survey is to give the current progress of vision-based deep learning HAR methods with the up-to-date study of literature.  ...  whether deep learning-based HAR methods are mapping with HAR datasets.  ... 
doi:10.1080/08839514.2022.2093705 fatcat:6on4g3sp3vaktnyyrk72k4mqta

Human Activity Recognition using Smartwatch and Smartphone: A Review on Methods, Applications, and Challenges

Rana Abdulrahman Lateef, Ayad Rodhan Abbas
2022 Iraqi Journal of Science  
and the advance of deep learning approaches.  ...  The literature is summarized from four aspects: sensors types, applications, Machine Learning (ML) and Deep Learning (DL) models, results and challenges.  ...  Employing HAR as an authentication system is still a challenge that is not exploited yet to cope the resultant high rate of error.  ... 
doi:10.24996/ijs.2022.63.1.34 fatcat:dznj4ouzirea5l34mngmptmhle
« Previous Showing results 1 — 15 out of 675 results