Blood pressure estimation throught function approximation
Periodic health control monitoring can assist us to minimize the effect of undiscovered chronic hypertension leading to several cardiovascular diseases. Blood Pressure (BP) is an important factor to be considered as it can provide fundamental information about personal health condition of a person....
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis-Master by Coursework |
| Language: | English |
| Published: |
Nanyang Technological University
2020
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/142929 |
| _version_ | 1826122585816956928 |
|---|---|
| author | Rajendran, Chandrakala |
| author2 | Saman S Abeysekera |
| author_facet | Saman S Abeysekera Rajendran, Chandrakala |
| author_sort | Rajendran, Chandrakala |
| collection | NTU |
| description | Periodic health control monitoring can assist us to minimize the effect of undiscovered chronic hypertension leading to several cardiovascular diseases. Blood Pressure (BP) is an important factor to be considered as it can provide fundamental information about personal health condition of a person. Hence it is necessary to implement a fast and easy BP measurement technique for the ease of patients anywhere, anytime. Photoplethysmography (PPG) based blood pressure measurement is more approachable due to its inexpensiveness and simplicity of measurement. It is observed that there exists a non-linear relationship between BP and time-domain features, obtained from PPG signal. This dissertation describes a new approach for estimating Systolic BP and Diastolic BP by non-invasively using PPG signal. This work involves prediction by Artificial Neural Network using modified four-element windkessel parameters obtained from photoplethysmography signal. Using MATLAB, the model is trained with input PPG parameters such as compliance of proximal C1 and distal C2 artery, total inertance L, Systolic time Ts to obtain the Systolic BP and Diastolic BP.
The performance of Artificial Neural Network is compared with Shallow Neural Network and Regression by Support Vector Machine by computing the mean square error, standard deviation, mean of each machine learning model. Compared with SVM-R and SNN, the approach of ANN has better accuracy with less mean squared error, standard deviation, and mean value of SBP and DBP. |
| first_indexed | 2024-10-01T05:50:34Z |
| format | Thesis-Master by Coursework |
| id | ntu-10356/142929 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T05:50:34Z |
| publishDate | 2020 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1429292023-07-04T15:02:49Z Blood pressure estimation throught function approximation Rajendran, Chandrakala Saman S Abeysekera School of Electrical and Electronic Engineering Esabeysekera@ntu.edu.sg Engineering::Electrical and electronic engineering Periodic health control monitoring can assist us to minimize the effect of undiscovered chronic hypertension leading to several cardiovascular diseases. Blood Pressure (BP) is an important factor to be considered as it can provide fundamental information about personal health condition of a person. Hence it is necessary to implement a fast and easy BP measurement technique for the ease of patients anywhere, anytime. Photoplethysmography (PPG) based blood pressure measurement is more approachable due to its inexpensiveness and simplicity of measurement. It is observed that there exists a non-linear relationship between BP and time-domain features, obtained from PPG signal. This dissertation describes a new approach for estimating Systolic BP and Diastolic BP by non-invasively using PPG signal. This work involves prediction by Artificial Neural Network using modified four-element windkessel parameters obtained from photoplethysmography signal. Using MATLAB, the model is trained with input PPG parameters such as compliance of proximal C1 and distal C2 artery, total inertance L, Systolic time Ts to obtain the Systolic BP and Diastolic BP. The performance of Artificial Neural Network is compared with Shallow Neural Network and Regression by Support Vector Machine by computing the mean square error, standard deviation, mean of each machine learning model. Compared with SVM-R and SNN, the approach of ANN has better accuracy with less mean squared error, standard deviation, and mean value of SBP and DBP. Master of Science (Signal Processing) 2020-07-14T01:40:11Z 2020-07-14T01:40:11Z 2020 Thesis-Master by Coursework https://hdl.handle.net/10356/142929 en application/pdf Nanyang Technological University |
| spellingShingle | Engineering::Electrical and electronic engineering Rajendran, Chandrakala Blood pressure estimation throught function approximation |
| title | Blood pressure estimation throught function approximation |
| title_full | Blood pressure estimation throught function approximation |
| title_fullStr | Blood pressure estimation throught function approximation |
| title_full_unstemmed | Blood pressure estimation throught function approximation |
| title_short | Blood pressure estimation throught function approximation |
| title_sort | blood pressure estimation throught function approximation |
| topic | Engineering::Electrical and electronic engineering |
| url | https://hdl.handle.net/10356/142929 |
| work_keys_str_mv | AT rajendranchandrakala bloodpressureestimationthroughtfunctionapproximation |