Human activity recognition: an approach 2D CNN-LSTM to sequential image representation and processing of inertial sensor data
The field of human activity recognition, abbreviated as HAR, benefits significantly from deep learning by addressing the complexity of human behavior and the vast volume of data produced by sensors. This work adopted the strategy of converting inertial data, such as accelerometer and gyroscope signa...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
AIMS Press
2024-11-01
|
| Series: | AIMS Bioengineering |
| Subjects: | |
| Online Access: | https://www.aimspress.com/article/doi/10.3934/bioeng.2024024 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832590265593888768 |
|---|---|
| author | Wallace Camacho Carlos Alessandro Copetti Luciano Bertini Leonard Barreto Moreira Otávio de Souza Martins Gomes |
| author_facet | Wallace Camacho Carlos Alessandro Copetti Luciano Bertini Leonard Barreto Moreira Otávio de Souza Martins Gomes |
| author_sort | Wallace Camacho Carlos |
| collection | DOAJ |
| description | The field of human activity recognition, abbreviated as HAR, benefits significantly from deep learning by addressing the complexity of human behavior and the vast volume of data produced by sensors. This work adopted the strategy of converting inertial data, such as accelerometer and gyroscope signals, into 2D images through recurrence plots. This approach facilitated the effective exploration of data input and neural network architectures. By utilizing the recent history of movements as input for the models, this study evaluated the impact of this methodology on HAR using two adapted architectures: 2D convolutional neural networks combined with long short-term memory layers (2D CNN-LSTM) and standalone 2D convolutional neural networks (2D CNN). Their performances were compared with other state-of-the-art deep learning models. The contributions of this study were threefold: the handling of input data, the development of the two network architectures for HAR, and the high accuracy achieved, ranging from 97% to 98%, on the public University of California, Irvine human activity recognition dataset (UCI-HAR). These results highlighted the benefit of incorporating temporal data to enhance accuracy in activity classification. |
| format | Article |
| id | doaj-art-7d471ad17e3c4c07ad9fd301cc962f3a |
| institution | Kabale University |
| issn | 2375-1495 |
| language | English |
| publishDate | 2024-11-01 |
| publisher | AIMS Press |
| record_format | Article |
| series | AIMS Bioengineering |
| spelling | doaj-art-7d471ad17e3c4c07ad9fd301cc962f3a2025-01-24T01:27:36ZengAIMS PressAIMS Bioengineering2375-14952024-11-0111452756010.3934/bioeng.2024024Human activity recognition: an approach 2D CNN-LSTM to sequential image representation and processing of inertial sensor dataWallace Camacho Carlos0Alessandro Copetti1Luciano Bertini2Leonard Barreto Moreira3Otávio de Souza Martins Gomes4Science and Technology Institute, Fluminense Federal University (ICT-UFF), Rio das Ostras, RJ, BrazilScience and Technology Institute, Fluminense Federal University (ICT-UFF), Rio das Ostras, RJ, BrazilInstitute of Systems Engineering and Information Technology, Federal University of Itajubá (IESTI-UNIFEI), Itajubá, MG, BrazilScience and Technology Institute, Fluminense Federal University (ICT-UFF), Rio das Ostras, RJ, BrazilInstitute of Systems Engineering and Information Technology, Federal University of Itajubá (IESTI-UNIFEI), Itajubá, MG, BrazilThe field of human activity recognition, abbreviated as HAR, benefits significantly from deep learning by addressing the complexity of human behavior and the vast volume of data produced by sensors. This work adopted the strategy of converting inertial data, such as accelerometer and gyroscope signals, into 2D images through recurrence plots. This approach facilitated the effective exploration of data input and neural network architectures. By utilizing the recent history of movements as input for the models, this study evaluated the impact of this methodology on HAR using two adapted architectures: 2D convolutional neural networks combined with long short-term memory layers (2D CNN-LSTM) and standalone 2D convolutional neural networks (2D CNN). Their performances were compared with other state-of-the-art deep learning models. The contributions of this study were threefold: the handling of input data, the development of the two network architectures for HAR, and the high accuracy achieved, ranging from 97% to 98%, on the public University of California, Irvine human activity recognition dataset (UCI-HAR). These results highlighted the benefit of incorporating temporal data to enhance accuracy in activity classification.https://www.aimspress.com/article/doi/10.3934/bioeng.2024024activity recognitionrecurrence plot2d-dlcnnlstm |
| spellingShingle | Wallace Camacho Carlos Alessandro Copetti Luciano Bertini Leonard Barreto Moreira Otávio de Souza Martins Gomes Human activity recognition: an approach 2D CNN-LSTM to sequential image representation and processing of inertial sensor data AIMS Bioengineering activity recognition recurrence plot 2d-dl cnn lstm |
| title | Human activity recognition: an approach 2D CNN-LSTM to sequential image representation and processing of inertial sensor data |
| title_full | Human activity recognition: an approach 2D CNN-LSTM to sequential image representation and processing of inertial sensor data |
| title_fullStr | Human activity recognition: an approach 2D CNN-LSTM to sequential image representation and processing of inertial sensor data |
| title_full_unstemmed | Human activity recognition: an approach 2D CNN-LSTM to sequential image representation and processing of inertial sensor data |
| title_short | Human activity recognition: an approach 2D CNN-LSTM to sequential image representation and processing of inertial sensor data |
| title_sort | human activity recognition an approach 2d cnn lstm to sequential image representation and processing of inertial sensor data |
| topic | activity recognition recurrence plot 2d-dl cnn lstm |
| url | https://www.aimspress.com/article/doi/10.3934/bioeng.2024024 |
| work_keys_str_mv | AT wallacecamachocarlos humanactivityrecognitionanapproach2dcnnlstmtosequentialimagerepresentationandprocessingofinertialsensordata AT alessandrocopetti humanactivityrecognitionanapproach2dcnnlstmtosequentialimagerepresentationandprocessingofinertialsensordata AT lucianobertini humanactivityrecognitionanapproach2dcnnlstmtosequentialimagerepresentationandprocessingofinertialsensordata AT leonardbarretomoreira humanactivityrecognitionanapproach2dcnnlstmtosequentialimagerepresentationandprocessingofinertialsensordata AT otaviodesouzamartinsgomes humanactivityrecognitionanapproach2dcnnlstmtosequentialimagerepresentationandprocessingofinertialsensordata |