Dissertation
Dissertation > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Text Processing > Text entry technology

Design and implementation of text input system based on steady state visual evoked potential

Author SuJingQiao
Tutor TangYu
School University of Electronic Science and Technology
Course Software Engineering
Keywords BCI SSVEP FFT DAUB4
CLC TP391.14
Type Master's thesis
Year 2012
Downloads 43
Quotes 0
Download Dissertation

The thesis presents a brain-computer interface (BCI) that can help users to inputword. The system is based on the steady-state visual evoked potential(SSVEP).Thesystem can adaptively increase or reduce its processing ability according to the specificspeed of generating data from various sources.We design a high precision software visual stimulator based on thehigh-resolution performance counter in the Windows environment. Considering thedifferent valid visual evoked frequency space of subject, the visual stimulator dividedinto5frequency points and5frequency points. After all, there are26letters to input,and the frequency interval can not be too small, we design to reuse every frequencypoints and make each of them correspond to more than one letter, and solve this basicproblem successfully.We design a high precision software visual stimulator based on thehigh-resolution performance counter in the Windows environment. Considering thedifferent valid visual evoked frequency space of subject, the visual stimulator dividedinto5frequency points and5frequency points. After all, there are26letters to input,and the frequency interval can not be too small, we design to reuse every frequencypoints and make each of them correspond to more than one letter, and solve this basicproblem successfully.Given the extreme importance of the feature extraction algorithm in this system,we implement the feature extraction algorithm including DAUB4filter algorithm andFFT of real functions, and describe in detail in this thesis. The feature extractionalgorithm isn’t simply integrated into the system, but exists as a DLL file. The reasonin this way is that we hope to pay long-term attention to the feature extractionalgorithm. Even those which are generally considered not suitable for the analysis ofEEG, maybe work efficiently for this specific BCI application. I wish to try anothervarious classes of algorithm to find the best feature extraction algorithm for SSVEP.

Related Dissertations
More Dissertations