| dc.identifier.uri | http://hdl.handle.net/11401/77438 | |
| dc.description.sponsorship | This work is sponsored by the Stony Brook University Graduate School in compliance with the requirements for completion of degree. | en_US |
| dc.format | Monograph | |
| dc.format.medium | Electronic Resource | en_US |
| dc.language.iso | en_US | |
| dc.publisher | The Graduate School, Stony Brook University: Stony Brook, NY. | |
| dc.type | Thesis | |
| dcterms.abstract | In the course of delivery, a fetus may suffer from oxygen deficiency due to the intensive pressure changes. Electronic fetal monitoring (EFM) system has been widely used in obstetrics, to provide continuous information to clinicians for making decisions and in preparing for delivery. There has been many efforts to build automated systems to analyze fetal heart rates (FHRs) and offer clinical supports. In this thesis, our goal is to introduce the most recent and popular machine learning method, deep learning, for FHR classification. We first introduce the preliminaries of FHR classification methods and the database used in our experiments. Then, the basics and unique characteristics of deep learning are discussed, in order to create foundation to understand our method. After that, we introduce 1-D convolutional layer to the models and select their parameters. Finally, we test the performance and generalization under three conditions. We build two models, which take the raw FHR and features extracted from FHR, respectively. The comparison between two models confirms the capability of neural network to exploit nonlinear features. We also apply data augmentation to the FHR database, which eliminates the unbalance of data set and the lack of sample size. It shows good performance of cross validation on augmented data set. The generalization of the models is tested on the original data set used to generate augmented data. Finally, we propose conjectures on the low true positive rate happened in the validation on original data set that is not used in generation. | |
| dcterms.available | 2017-09-20T16:52:41Z | |
| dcterms.contributor | Djurić, Petar M | en_US |
| dcterms.contributor | Bugallo, Mónica F. | en_US |
| dcterms.creator | CHEN, XUAN | |
| dcterms.dateAccepted | 2017-09-20T16:52:41Z | |
| dcterms.dateSubmitted | 2017-09-20T16:52:41Z | |
| dcterms.description | Department of Electrical Engineering | en_US |
| dcterms.extent | 67 pg. | en_US |
| dcterms.format | Monograph | |
| dcterms.format | Application/PDF | en_US |
| dcterms.identifier | http://hdl.handle.net/11401/77438 | |
| dcterms.issued | 2016-12-01 | |
| dcterms.language | en_US | |
| dcterms.provenance | Made available in DSpace on 2017-09-20T16:52:41Z (GMT). No. of bitstreams: 1
CHEN_grad.sunysb_0771M_13165.pdf: 1294523 bytes, checksum: 343ada4d4a07737778722277597353e1 (MD5)
Previous issue date: 1 | en |
| dcterms.publisher | The Graduate School, Stony Brook University: Stony Brook, NY. | |
| dcterms.subject | Electrical engineering | |
| dcterms.title | Classification of Fetal Heart Rate Signals by Deep Learning | |
| dcterms.type | Thesis | |
