You are here

Kutatószeminárium

Félév: 
2019/20 I. félév
Helyszín: 
Árpád tér 2. II. em. 220. sz.
Dátum: 
2019-12-10
Időpont: 
15:15 16:00
Előadó: 
Amin Honarmandi
Cím: 
Unspoken Speech
Absztrakt: 

Unspoken speech detection from brain activity is a work out field of investigation, which has penetrated significant progress during the two past decade. The objective of this study is to investigate the use of common features in different classifiers that might be related to find best discriminate EEG responses for each imagine and verbal phonemic and single word prompts on a trial basis. First we identify noise and artifacts, then discard this signals with EEGLAB software. After that, decompose signals to intrinsic mode functions which allow to decompose to different band frequency, we then apply CSP for extract features, finally we apply classifications method. Common features in different classifiers are overall significant, instead we use huge features for classifying we can use common features with relative better accuracy for all classifiers.