Thesis
Signal processing for detection and classification
Washington State University
Master of Science (MS), Washington State University
05/2020
DOI:
https://doi.org/10.7273/000004092
Handle:
https://hdl.handle.net/2376/124755
Abstract
With the rise of the internet and digital information technology, various signals are flooding our world. As a result, detection and classification of signals have become very important. Traditionally, fixed features were been used to detect and classify signals. However, complexity and quantity of signals are extremely increased in recent years; as a result, analysis through artificial intelligence is inevitable. In the thesis, a new signal processing design for detecting seizure from Electroencephalogram (EEG) recorded from healthy people and epileptic patients has been introduced. Discrete wavelet transform (DWT) and neural network are used to analyze EEG signal. Five groups of EEG signal samples, in which three groups from patients and two groups from healthy subjects, have been tested in the study. With information from both time domain provided by DWT and a high-performance neural network model, the detection system shows a high accuracy on detecting ictal seizures signal, which was approximately 95%. Furthermore, five specific clinical questions have been addressed, and the model for each question has been trained independently. Results for these questions have been judged by calculating F1 score for binary classification or micro- macro- F1 score for multi-class classification. The results showed that the size of samples is a key point toward training a high-performance neural network model.
Metrics
5 File views/ downloads
37 Record Views
Details
- Title
- Signal processing for detection and classification
- Creators
- Haotian Yang
- Contributors
- Praveen K Sekhar (Advisor) - Washington State University, School of Engineering and Computer Science (VANC)
- Awarding Institution
- Washington State University
- Academic Unit
- School of Engineering and Computer Science (VANC)
- Theses and Dissertations
- Master of Science (MS), Washington State University
- Publisher
- Washington State University
- Identifiers
- 99900890787401842
- Language
- English
- Resource Type
- Thesis