The pre-impact Fall Detection using the quantization based on ResNet...World Congress of Gerontechnology, October 22-26, 2022, Daegu, South Korea.
Purpose Fall is one of the major health risks for older people. Hip fractures are considered the most dangerous among various injuries, as they can reduce mobility and lead to many complications, and even death (Hagen, G et al., 2020). Some researchers tried to protect the user using a wearable airb...
| Publicado en: | Gerontechnology Vol. 21; pp. 1 - 2 |
|---|---|
| Autores principales: | , , , , , |
| Formato: | abstract proceedings research Journal Article |
| Publicado: |
International Society for Gerontechnology
Oct2022
|
| Acceso en línea: | Ver este registro en EBSCOhost |
| fields | @attributes: recordID: 1 pdfLink: plink: https://search.ebscohost.com/login.aspx?direct=true&db=ccm&AN=161396244&site=ehost-live header: @attributes: shortDbName: ccm uiTerm: 161396244 longDbName: CINAHL Complete uiTag: AN controlInfo: bkinfo: dissinfo: jinfo: jid: 15691101 904R jtl: Gerontechnology issn: 15691101 maglogo: N pubinfo: dt: Oct2022 vid: 21 pid: 54298 pub: International Society for Gerontechnology artinfo: ui: 161396244 161396244 161396244 10.4017/gt.2022.21.s.675.pp1 161396244 ppf: 1 ppct: 1 formats: tig: atl: The pre-impact Fall Detection using the quantization based on ResNet...World Congress of Gerontechnology, October 22-26, 2022, Daegu, South Korea. aug: au: Koo, B. M. Kim, J. M. Yang, S. M. Lee, S. H. Hong, M. H. Kim, Y. H. affil: Department of Biomedical Engineering and Institute of Medical Engineering, Yonsei University, Republic of Korea sug: subj: Accidental Falls Prevention and Control Deep Learning Algorithms Neural Networks (Computer) Congresses and Conferences South Korea South Korea ab: Purpose Fall is one of the major health risks for older people. Hip fractures are considered the most dangerous among various injuries, as they can reduce mobility and lead to many complications, and even death (Hagen, G et al., 2020). Some researchers tried to protect the user using a wearable airbag based on the threshold-based algorithm (Jung, H et al., 2021). The more accurate algorithm was developed using the deep learning methods, but it was heavy and needed GPU or PC (Yu, X et al., 2021). This study was focused on developing the lightweight deep learning algorithm to detect pre-impact falls for targeting edge devices. Method The KFall public dataset was used in this study. The experimental protocol consisted of 15 fall movements and 21 activities of daily living (ADLs). The 9-axis IMU sensor data with 100 Hz sampling frequency were acquired from the waists of 32 young male subjects. Data of 26 subjects were used to train the deep learning models and those of remained 6 subjects were used to test the models. We developed the fall detection algorithm based on ResNet method. It consisted of 2 convolution layers, 4 convolution blocks, 3 identity blocks, 1 average pooling layer, and 1 SoftMax layer. We removed one by one except for the factors that affect the output size. ResNet14 model was transformed to a flat buffer type and two quantization techniques were applied using TensorFlow Lite. Results and Discussion Table 1 showed the performance of ResNet algorithms. When continuously reducing the layers, demanded memory decreased, but there was no dramatic change in accuracy. It suggested that our model was sufficiently deep and additional identity blocks were not needed. Table 2 showed the performance of ResNet14 models according to quantization. The integer quantization showed the smallest size of memory but the worst accuracy. The float16 quantization showed the bigger size of memory than the integer quantization, but the accuracy was maintained. As a result, the float16 quantization showed 98% of accuracy with 539 ±278 ms of sufficient lead time and 0.104 MB of memory size. pubtype: Academic Journal doctype: abstract proceedings research Journal Article ougenre: Article language: English refInfo: holdings: @attributes: islocal: N |
|---|