Prediction of urinal flowrate by using urinary acoustic signals based on LSTM neural network
- Title
- Prediction of urinal flowrate by using urinary acoustic signals based on LSTM neural network
- Author
- 전진용
- Issue Date
- 2020-08
- Publisher
- International Institute of Noise Control Engineering (I-INCE)
- Citation
- INTER-NOISE and NOISE-CON Congress and Conference Proceedings, page. 4995-5868
- Abstract
- To measure urinary patterns of the patients with urination disorders anywhere and to predict various symptoms of lower urinary tract ultimately, this study was conducted to produce a device that was capable of analysing and classifying the urinary acoustic signals. Assuming the loudness representing the magnitude of the urinary sound, the roughness representing the signal change pattern as inputs and the flowrate measured through the uroflowmetry test as output, the LSTM model for time series prediction was trained and the system was constructed. The predicted values using LSTM model and the measured result are quite similar. In addition, based on the predicted time series velocity results, it is possible to check the health status of the urinary system through the LSTM classification model and the abnormal symptoms of the lower urinary tract can be suspected. Therefore, it is shown that the LSTM prediction and classification model proposed in this study can be applied to the clinic for predicting various symptoms with urinary acoustic signals.
- URI
- https://www.ingentaconnect.com/content/ince/incecp/2020/00000261/00000001/art00093https://repository.hanyang.ac.kr/handle/20.500.11754/169866
- ISSN
- 0736-2935
- Appears in Collections:
- COLLEGE OF ENGINEERING[S](공과대학) > ARCHITECTURAL ENGINEERING(건축공학부) > Articles
- Files in This Item:
There are no files associated with this item.
- Export
- RIS (EndNote)
- XLS (Excel)
- XML