2021
DOI: 10.1109/access.2021.3070646
|View full text |Cite
|
Sign up to set email alerts
|

Boosting Inertial-Based Human Activity Recognition With Transformers

Abstract: Activity recognition problems such as human activity recognition and smartphone location recognition can improve the accuracy of different navigation or healthcare tasks, which rely solely on inertial sensors. Current learning-based approaches for activity recognition from inertial data employ convolutional neural networks or long short term memory architectures. Recently, Transformers were shown to outperform these architectures for sequence analysis tasks. This work presents an activity recognition model bas… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
2
1

Citation Types

1
28
0

Year Published

2022
2022
2024
2024

Publication Types

Select...
4
3
1

Relationship

1
7

Authors

Journals

citations
Cited by 54 publications
(29 citation statements)
references
References 30 publications
(37 reference statements)
1
28
0
Order By: Relevance
“…To evaluate the performance of the proposed model, the transformer-based-model and the 1D-CNN model, proposed by Shavit [ 22 ], were tested together. For the direct comparison of the conformer and transformer structures, the remaining structures, except for the conformer and transformer structures, of the two models are made completely identical.…”
Section: Resultsmentioning
confidence: 99%
See 2 more Smart Citations
“…To evaluate the performance of the proposed model, the transformer-based-model and the 1D-CNN model, proposed by Shavit [ 22 ], were tested together. For the direct comparison of the conformer and transformer structures, the remaining structures, except for the conformer and transformer structures, of the two models are made completely identical.…”
Section: Resultsmentioning
confidence: 99%
“…The transformer model is known to have good computational efficiency while extracting temporal dependency similar to the RNN model [ 17 ] and has been recently introduced into IMU-based HAR research. In Shavit’s study [ 22 ], the transformer encoder structure was used for the first time in HAR research. Input data were converted into a latent sequence embedding layer and then passed through a transformer encoder.…”
Section: Related Workmentioning
confidence: 99%
See 1 more Smart Citation
“…Human Activity Recognition. In recent years, a large body of research for wearable-based human activity recognition is dedicated to learning discriminative features by leveraging various deep neural networks, including convolutional networks, residual networks, autoencoders and Transformers [7,22,26,33,41]. These models are accurate when the number of training instances is sufficiently large, yet such performance is not guaranteed when the labels are scarce.…”
Section: Related Workmentioning
confidence: 99%
“…In [23][24][25], a deep learning approach is used for robot indoor navigation. Human activity recognition [26][27][28] and smartphone location recognition (SLR) [29] algorithms based on ML/DL were shown to improve the accuracy of PDR by using it as a prior [30,31]. SLR was also shown to improve the performance adaptive attitude and heading reference system (AHRS) [32].…”
Section: Introductionmentioning
confidence: 99%