izpis_h1_title_alt

Trajno napovedovanje krvnega tlaka iz signala PPG
Slapničar, Gašper (Author), Kukar, Matjaž (Mentor) More about this mentor... This link opens in a new window, Luštrek, Mitja (Co-mentor)

.pdfPDF - Presentation file, Download (2,10 MB)
MD5: 6789C21EC9FAA8CBD3EE06D69A2D1AEA

Abstract
Krvni tlak je pomemben pokazatelj hipertenzije. Razvili smo sistem, ki krvni tlak ocenjuje iz fotopletizmograma (PPG), kakršen je že vgrajen v večino modernih senzorskih zapestnic. Zaradi šuma in motenj, ki se v signalu PPG pojavijo kot posledica uporabe zapestnice, smo razvili metodo čiščenja in segmentiranja signala PPG na cikle. Nato smo izračunali množico značilk, ki smo jih uporabili v regresijskih modelih. Sistem smo izboljšali z uporabo algoritma RReliefF za izbor najboljših značilk in z uporabo dela podatkov vsake osebe za učenje personaliziranih napovednih modelov. Sistem smo vrednotili na dveh podatkovnih množicah, eni iz kliničnega okolja in drugi zbrani med rutinskimi dnevnimi aktivnostmi posameznikov. V poizkusu, kjer model vsakič naučimo na vseh osebah razen eni in ga nato testiramo na izpuščeni osebi, smo z uporabo klinične množice (podatkovna baza MIMIC) dosegli najnižjo povprečno absolutno napako (MAE) 5,61 mmHg za sistolični in 3,82 mmHg za diastolični krvni tlak, oboje pri največji stopnji personalizacije. Za množico, zbrano med rutinskimi dnevnimi aktivnostmi, smo dosegli najnižjo MAE 8.40 mmHg za sistolični in 4.20 mmHg za diastolični krvni tlak, ponovno pri največji stopnji personalizacije. Najbolje sta se obnesla algoritma globoka regresija in ``naključni gozd''. Rezultati skoraj dosegajo zahteve dveh glavnih standardov za ocenjevanje krvenga tlaka.

Language:English
Keywords:krvni tlak, fotopletizmografija, strojno učenje, regresija, obdelava signalov
Work type:Master's thesis/paper (mb22)
Organization:FRI - Faculty of computer and information science
Year:2018
Views:903
Downloads:425
Metadata:XML RDF-CHPDL DC-XML DC-RDF
 
Average score:(0 votes)
Your score:Voting is allowed only to logged in users.
:
Share:AddThis
AddThis uses cookies that require your consent. Edit consent...

Secondary language

Language:Slovenian
Title:Continuous blood pressure estimation from PPG signal
Abstract:
Blood pressure (BP) is an indicator of hypertension. We developed a system in which photoplethysmogram (PPG), which is commonly integrated in modern wearables, is used to continuously estimate BP. A preprocessing module was developed and used for cleaning the PPG signal of noise and artefacts, and segmenting it into cycles. A set of features describing the PPG signal was then computed to be used in regression models. The RReliefF algorithm was used to select a subset of relevant features and personalization of the models was considered to further improve the performance of the models. The approach was validated using two distinct datasets, one from a hospital environment, and the other collected during every-day activities. Using the clinical dataset (MIMIC database), the best achieved mean absolute errors (MAE) in a leave-one-subject-out (LOSO) experiment were 5.61 mmHg for systolic and 3.82 mmHg for diastolic BP, at maximum personalization. For everyday-life dataset, the lowest errors were 8.40 mmHg for systolic and 4.20 mmHg for diastolic BP. Deep learning regression and Random Forest algorithm achieved the best results. Our results borderline meet the requirements of the two most well-established standards for BP estimation devices.

Keywords:blood pressure, photoplethysmography, machine learning, regression, signal processing

Similar documents

Similar works from RUL:
Similar works from other Slovenian collections:

Comments

Leave comment

You have to log in to leave a comment.

Comments (0)
0 - 0 / 0
 
There are no comments!

Back