Comparison of End-to-End Neural Network Architectures and Data Augmentation Methods for Automatic Infant Motility Assessment Using Wearable Sensors
Infant motility assessment using intelligent wearables is a promising new approach for assessment of infant neurophysiological development, and where efficient signal analysis plays a central role. This study investigates the use of different end-to-end neural network architectures for processing in...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-04-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/23/7/3773 |
| _version_ | 1797606977984004096 |
|---|---|
| author | Manu Airaksinen Sampsa Vanhatalo Okko Räsänen |
| author_facet | Manu Airaksinen Sampsa Vanhatalo Okko Räsänen |
| author_sort | Manu Airaksinen |
| collection | DOAJ |
| description | Infant motility assessment using intelligent wearables is a promising new approach for assessment of infant neurophysiological development, and where efficient signal analysis plays a central role. This study investigates the use of different end-to-end neural network architectures for processing infant motility data from wearable sensors. We focus on the performance and computational burden of alternative sensor encoder and time series modeling modules and their combinations. In addition, we explore the benefits of data augmentation methods in ideal and nonideal recording conditions. The experiments are conducted using a dataset of multisensor movement recordings from 7-month-old infants, as captured by a recently proposed smart jumpsuit for infant motility assessment. Our results indicate that the choice of the encoder module has a major impact on classifier performance. For sensor encoders, the best performance was obtained with parallel two-dimensional convolutions for intrasensor channel fusion with shared weights for all sensors. The results also indicate that a relatively compact feature representation is obtainable for within-sensor feature extraction without a drastic loss to classifier performance. Comparison of time series models revealed that feedforward dilated convolutions with residual and skip connections outperformed all recurrent neural network (RNN)-based models in performance, training time, and training stability. The experiments also indicate that data augmentation improves model robustness in simulated packet loss or sensor dropout scenarios. In particular, signal- and sensor-dropout-based augmentation strategies provided considerable boosts to performance without negatively affecting the baseline performance. Overall, the results provide tangible suggestions on how to optimize end-to-end neural network training for multichannel movement sensor data. |
| first_indexed | 2024-03-11T05:23:57Z |
| format | Article |
| id | doaj.art-77a432766a004c0e94b759dabe9e5906 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-11T05:23:57Z |
| publishDate | 2023-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-77a432766a004c0e94b759dabe9e59062023-11-17T17:37:34ZengMDPI AGSensors1424-82202023-04-01237377310.3390/s23073773Comparison of End-to-End Neural Network Architectures and Data Augmentation Methods for Automatic Infant Motility Assessment Using Wearable SensorsManu Airaksinen0Sampsa Vanhatalo1Okko Räsänen2BABA Center, Pediatric Research Center, Children’s Hospital, Helsinki University Hospital and University of Helsinki, 00290 Helsinki, FinlandUnit of Computing Sciences, Tampere University, 33720 Tampere, FinlandUnit of Computing Sciences, Tampere University, 33720 Tampere, FinlandInfant motility assessment using intelligent wearables is a promising new approach for assessment of infant neurophysiological development, and where efficient signal analysis plays a central role. This study investigates the use of different end-to-end neural network architectures for processing infant motility data from wearable sensors. We focus on the performance and computational burden of alternative sensor encoder and time series modeling modules and their combinations. In addition, we explore the benefits of data augmentation methods in ideal and nonideal recording conditions. The experiments are conducted using a dataset of multisensor movement recordings from 7-month-old infants, as captured by a recently proposed smart jumpsuit for infant motility assessment. Our results indicate that the choice of the encoder module has a major impact on classifier performance. For sensor encoders, the best performance was obtained with parallel two-dimensional convolutions for intrasensor channel fusion with shared weights for all sensors. The results also indicate that a relatively compact feature representation is obtainable for within-sensor feature extraction without a drastic loss to classifier performance. Comparison of time series models revealed that feedforward dilated convolutions with residual and skip connections outperformed all recurrent neural network (RNN)-based models in performance, training time, and training stability. The experiments also indicate that data augmentation improves model robustness in simulated packet loss or sensor dropout scenarios. In particular, signal- and sensor-dropout-based augmentation strategies provided considerable boosts to performance without negatively affecting the baseline performance. Overall, the results provide tangible suggestions on how to optimize end-to-end neural network training for multichannel movement sensor data.https://www.mdpi.com/1424-8220/23/7/3773human activity recognitionclassifier architectureswearable technologyinfant motility |
| spellingShingle | Manu Airaksinen Sampsa Vanhatalo Okko Räsänen Comparison of End-to-End Neural Network Architectures and Data Augmentation Methods for Automatic Infant Motility Assessment Using Wearable Sensors Sensors human activity recognition classifier architectures wearable technology infant motility |
| title | Comparison of End-to-End Neural Network Architectures and Data Augmentation Methods for Automatic Infant Motility Assessment Using Wearable Sensors |
| title_full | Comparison of End-to-End Neural Network Architectures and Data Augmentation Methods for Automatic Infant Motility Assessment Using Wearable Sensors |
| title_fullStr | Comparison of End-to-End Neural Network Architectures and Data Augmentation Methods for Automatic Infant Motility Assessment Using Wearable Sensors |
| title_full_unstemmed | Comparison of End-to-End Neural Network Architectures and Data Augmentation Methods for Automatic Infant Motility Assessment Using Wearable Sensors |
| title_short | Comparison of End-to-End Neural Network Architectures and Data Augmentation Methods for Automatic Infant Motility Assessment Using Wearable Sensors |
| title_sort | comparison of end to end neural network architectures and data augmentation methods for automatic infant motility assessment using wearable sensors |
| topic | human activity recognition classifier architectures wearable technology infant motility |
| url | https://www.mdpi.com/1424-8220/23/7/3773 |
| work_keys_str_mv | AT manuairaksinen comparisonofendtoendneuralnetworkarchitecturesanddataaugmentationmethodsforautomaticinfantmotilityassessmentusingwearablesensors AT sampsavanhatalo comparisonofendtoendneuralnetworkarchitecturesanddataaugmentationmethodsforautomaticinfantmotilityassessmentusingwearablesensors AT okkorasanen comparisonofendtoendneuralnetworkarchitecturesanddataaugmentationmethodsforautomaticinfantmotilityassessmentusingwearablesensors |