I have .ov file of P300 signals acquired using the example scenario of P300. Now, I want to use it for classification, but there is no label in it. I have converted the .ov file to .csv and the file has the following format:
Time (s);F4;T7;F8;T8;FP1;F7;FP2;F3;O1;O2;C4;P4;P3;Cz;Pz;C3;Sampling Rate
which shows that there is no label here. How can I do classification here. Infact I need to use some deep learning classification in tensorflow or Matlab as the classifiers in OpenViBE only contain LDA and SVM. Any help is appreciated.
Labels in P300 signals
-
- Posts: 775
- Joined: Tue Dec 04, 2012 3:53 pm
- Location: INRIA Rennes, FRANCE
Re: Labels in P300 signals
There should be, but the CSV conversion possibly loses the labels, as the ov exporter cannot put both in the same file. Openvibe .ov files can include multiple streams, but to get access to all of them, you need to have corresponding outputs defined in the Acquisition Client box and then route them to somewhere. In the case of csv, you need to use two separate csv boxes and files, one for signal, one for labels (called stimulations) in openvibe.
Happy hacking,
Jussi
Happy hacking,
Jussi
Re: Labels in P300 signals
Thanks for your reply. Now, I have used two CSV writers, and the stimulations obtained are:
Time (s);Identifier;Duration
1.9942382813e+001;32769;0.0000000000e+000
however, it does not make sense to me. I am doing a P300 experiment and want to do some classifications out of the openViBE for which the data should have labels. Also, how to merge this with 2nd csv file having signals data as stimulations file has far less number of rows. Any help will be appreciated.
Time (s);Identifier;Duration
1.9942382813e+001;32769;0.0000000000e+000
however, it does not make sense to me. I am doing a P300 experiment and want to do some classifications out of the openViBE for which the data should have labels. Also, how to merge this with 2nd csv file having signals data as stimulations file has far less number of rows. Any help will be appreciated.
-
- Posts: 775
- Joined: Tue Dec 04, 2012 3:53 pm
- Location: INRIA Rennes, FRANCE
Re: Labels in P300 signals
Both files will contain the timestamp, you need to match the timestamp of the event of interest in the stimulation matrix to the corresponding time in the signal file, and then extract a segment of the signal starting from that point (or a bit earlier if you need). If both files start from 0 (as is usually the case), basically you use the samplingRate*eventTimeStamp to get the indexes to the signal matrix.
For description of the stimulation codes, please see
http://openvibe.inria.fr/stimulation-codes/
Cheers,
Jussi
For description of the stimulation codes, please see
http://openvibe.inria.fr/stimulation-codes/
Cheers,
Jussi