EEG Signal Analysis: Color Recognition Using Support Vector Machine

by Jorge Heredia

Institution: California State University – Northridge
Year: 2015
Keywords: Color Recognition; Dissertations, Academic  – CSUN  – Engineering  – Electrical and Computer Engineering.
Posted: 02/05/2017
Record ID: 2134582
Full text PDF: http://hdl.handle.net/10211.3/143767


This project is a biomedical application of pattern recognition. The objective of this project is to be able to accurately predict which of two colors, red or blue, is seen by the subject by analyzing the EEG signals and patterns present for each color. Techniques utilized in this paper include the Surface Laplacian spatial filter, power line rejection notch filter, spectral analysis using the FFT, PCA for feature extraction, and SVM for machine learning. The data acquisition device used for this project was the Emotiv EPOC headset, which follows the international 10-20 channel placement locations and has a sampling rate of 128Hz. The project was successfully able to predict the colors. Advisors/Committee Members: Hang, Xiyi (advisor), Katz, Sharlene (committee member).