OpenViBE forum

Hi all,



I wonder if the P300 Speller training scenario can handle an input of several .ov files.



I suppose that training the classifier with one single file (i.e., one single word) is not enough for reaching high classifier performances. Hence, owing to the fact that using a word with a huge number of characters would tire the user, it would be suitable to train the LDA classifier with several files.

If this is not possible, I wonder if the multiple .ov files can be merged somehow in order to “fake” a larger word to train.

Thanks in advance.

Hi, the following thread might be of use,

viewtopic.php?f=17&t=9782&hilit=feature

If you're combining many datasets of the same user on the same day and same cap setup for P300, that could help (in the post its about motor imagery + multiple sessions/users afaik).


Cheers,
Jussi