The Performance of Post-Fall Detection Using the Cross-Dataset: Feature Vectors, Classifiers and Processing Conditions
In this study, algorithms to detect post-falls were evaluated using the cross-dataset according to feature vectors (time-series and discrete data), classifiers (ANN and SVM), and four different processing conditions (normalization, equalization, increase in the number of training data, and additiona...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2021-07-01
|
Series: | Sensors |
Subjects: | |
Online Access: | https://www.mdpi.com/1424-8220/21/14/4638 |
_version_ | 1797526144886505472 |
---|---|
author | Bummo Koo Jongman Kim Yejin Nam Youngho Kim |
author_facet | Bummo Koo Jongman Kim Yejin Nam Youngho Kim |
author_sort | Bummo Koo |
collection | DOAJ |
description | In this study, algorithms to detect post-falls were evaluated using the cross-dataset according to feature vectors (time-series and discrete data), classifiers (ANN and SVM), and four different processing conditions (normalization, equalization, increase in the number of training data, and additional training with external data). Three-axis acceleration and angular velocity data were obtained from 30 healthy male subjects by attaching an IMU to the middle of the left and right anterior superior iliac spines (ASIS). Internal and external tests were performed using our lab dataset and SisFall public dataset, respectively. The results showed that ANN and SVM were suitable for the time-series and discrete data, respectively. The classification performance generally decreased, and thus, specific feature vectors from the raw data were necessary when untrained motions were tested using a public dataset. Normalization made SVM and ANN more and less effective, respectively. Equalization increased the sensitivity, even though it did not improve the overall performance. The increase in the number of training data also improved the classification performance. Machine learning was vulnerable to untrained motions, and data of various movements were needed for the training. |
first_indexed | 2024-03-10T09:25:03Z |
format | Article |
id | doaj.art-c3cf42fb213b428cbc56f0125e48fda6 |
institution | Directory Open Access Journal |
issn | 1424-8220 |
language | English |
last_indexed | 2024-03-10T09:25:03Z |
publishDate | 2021-07-01 |
publisher | MDPI AG |
record_format | Article |
series | Sensors |
spelling | doaj.art-c3cf42fb213b428cbc56f0125e48fda62023-11-22T04:53:52ZengMDPI AGSensors1424-82202021-07-012114463810.3390/s21144638The Performance of Post-Fall Detection Using the Cross-Dataset: Feature Vectors, Classifiers and Processing ConditionsBummo Koo0Jongman Kim1Yejin Nam2Youngho Kim3Department of Biomedical Engineering, Yonsei University, Wonju 26493, KoreaDepartment of Biomedical Engineering, Yonsei University, Wonju 26493, KoreaDepartment of Biomedical Engineering, Yonsei University, Wonju 26493, KoreaDepartment of Biomedical Engineering, Yonsei University, Wonju 26493, KoreaIn this study, algorithms to detect post-falls were evaluated using the cross-dataset according to feature vectors (time-series and discrete data), classifiers (ANN and SVM), and four different processing conditions (normalization, equalization, increase in the number of training data, and additional training with external data). Three-axis acceleration and angular velocity data were obtained from 30 healthy male subjects by attaching an IMU to the middle of the left and right anterior superior iliac spines (ASIS). Internal and external tests were performed using our lab dataset and SisFall public dataset, respectively. The results showed that ANN and SVM were suitable for the time-series and discrete data, respectively. The classification performance generally decreased, and thus, specific feature vectors from the raw data were necessary when untrained motions were tested using a public dataset. Normalization made SVM and ANN more and less effective, respectively. Equalization increased the sensitivity, even though it did not improve the overall performance. The increase in the number of training data also improved the classification performance. Machine learning was vulnerable to untrained motions, and data of various movements were needed for the training.https://www.mdpi.com/1424-8220/21/14/4638fall detectionartificial neural networksupport vector machinecross-dataset |
spellingShingle | Bummo Koo Jongman Kim Yejin Nam Youngho Kim The Performance of Post-Fall Detection Using the Cross-Dataset: Feature Vectors, Classifiers and Processing Conditions Sensors fall detection artificial neural network support vector machine cross-dataset |
title | The Performance of Post-Fall Detection Using the Cross-Dataset: Feature Vectors, Classifiers and Processing Conditions |
title_full | The Performance of Post-Fall Detection Using the Cross-Dataset: Feature Vectors, Classifiers and Processing Conditions |
title_fullStr | The Performance of Post-Fall Detection Using the Cross-Dataset: Feature Vectors, Classifiers and Processing Conditions |
title_full_unstemmed | The Performance of Post-Fall Detection Using the Cross-Dataset: Feature Vectors, Classifiers and Processing Conditions |
title_short | The Performance of Post-Fall Detection Using the Cross-Dataset: Feature Vectors, Classifiers and Processing Conditions |
title_sort | performance of post fall detection using the cross dataset feature vectors classifiers and processing conditions |
topic | fall detection artificial neural network support vector machine cross-dataset |
url | https://www.mdpi.com/1424-8220/21/14/4638 |
work_keys_str_mv | AT bummokoo theperformanceofpostfalldetectionusingthecrossdatasetfeaturevectorsclassifiersandprocessingconditions AT jongmankim theperformanceofpostfalldetectionusingthecrossdatasetfeaturevectorsclassifiersandprocessingconditions AT yejinnam theperformanceofpostfalldetectionusingthecrossdatasetfeaturevectorsclassifiersandprocessingconditions AT younghokim theperformanceofpostfalldetectionusingthecrossdatasetfeaturevectorsclassifiersandprocessingconditions AT bummokoo performanceofpostfalldetectionusingthecrossdatasetfeaturevectorsclassifiersandprocessingconditions AT jongmankim performanceofpostfalldetectionusingthecrossdatasetfeaturevectorsclassifiersandprocessingconditions AT yejinnam performanceofpostfalldetectionusingthecrossdatasetfeaturevectorsclassifiersandprocessingconditions AT younghokim performanceofpostfalldetectionusingthecrossdatasetfeaturevectorsclassifiersandprocessingconditions |