Search found 37 matches
- Fri Sep 28, 2018 6:04 pm
- Forum: Scenarios and BCI design
- Topic: How to make the motor imagery CSP work ?
- Replies: 3
- Views: 5011
Re: How to make the motor imagery CSP work ?
Maybe enlarging the number of trials (collecting more signals) for each class could help it to train/generalize/work better online ? in mi-csp-1-acquisition.xml: <Setting> <TypeIdentifier>(0x007deef9, 0x2f3e95c6)</TypeIdentifier> <Name>Number of Trials for Each Class</Name> <DefaultValue>20</Default...
- Thu May 05, 2016 5:58 pm
- Forum: Box and application development
- Topic: SpecCSP algorithm and companion SpatioSpectral filter boxes
- Replies: 1
- Views: 9105
Re: SpecCSP algorithm and companion SpatioSpectral filter bo
...and the last source file (the forum is limited to 3 attached files par post)
- Thu May 05, 2016 5:57 pm
- Forum: Box and application development
- Topic: SpecCSP algorithm and companion SpatioSpectral filter boxes
- Replies: 1
- Views: 9105
SpecCSP algorithm and companion SpatioSpectral filter boxes
Hi Openvibers, Here is the last version (v0.9) of the SpecCSP algorithm, which now estimates ARMA models (or IIR/FIR filters) of the spectral weightenning sets generated by the algorithm (via a Yule-Walker procedure) so as to apply the spectral weights via the CApplyTemporalFilter methods available ...
- Thu May 05, 2016 4:04 pm
- Forum: Designer
- Topic: Available imagery datasets on OV
- Replies: 0
- Views: 3855
Available imagery datasets on OV
Hi Openvibers, There is one imagery dataset bundled with OV (http://openvibe.inria.fr/datasets/), and a dedicated reader for the BCI competition IIIb, but - would you have other imagery datasets compatible with the OV generic reader ? (and sharable) - or other scenario readers specifically developpe...
- Tue Apr 26, 2016 3:41 pm
- Forum: Box and application development
- Topic: PSD estimate through Welch's method
- Replies: 2
- Views: 10678
Re: PSD estimate through Welch's method
Hi Matt, I had the same idea while using the itpp library for the SpecCSP algorithm :) the Welch estimator is already implemented (see ITPP_EXPORT vec itpp::spectrum() at http://itpp.sourceforge.net/4.3.1/group__sigproc.html), and some other PSD estimators could also be among the options of such a b...
- Mon Apr 25, 2016 5:40 pm
- Forum: Box and application development
- Topic: Spectrally weighted CSP OV algorithm box: help to validate ?
- Replies: 6
- Views: 15198
Re: Spectrally weighted CSP OV algorithm box: help to valida
Thank you Jussi ! The generated spatial heatmaps are pretty much consistent between each others and with the spatial filters generated by the other OV spatial filter trainers. But I still want to validate it step by step (the ones numbered in the code) along with Matlab. The boost library is a pain ...
- Fri Apr 15, 2016 5:59 pm
- Forum: Box and application development
- Topic: Spectrally weighted CSP OV algorithm box: help to validate ?
- Replies: 6
- Views: 15198
Re: Spectrally weighted CSP OV algorithm box: help to valida
Hi there, Here is my last version (version 0.6) of the SpecCSP algorithm. - Data structures (2d, 3d and 4d real or complex matrix) migrated toward boost::multi_array - Dual CSP mode added: CSP1 is based on the eigendecomposition eig(inv(R2)*R1) while CSP2 is eig(White*R1*White') where White is the w...
- Thu Apr 14, 2016 12:39 pm
- Forum: Box and application development
- Topic: Spectrally weighted CSP OV algorithm box: help to validate ?
- Replies: 6
- Views: 15198
Re: Spectrally weighted CSP OV algorithm box: help to valida
Hi OVibers, There is an error for the eq. 4 of the implementation, it was: alpha_opt{c}(k) = max( 0 , ( mu_s{c}(k) - mu_s{3-c}(k) ) / ( var_s{1}(k) + var_s{2}(k) ) ); ... and should be: alpha_opt{c}(k) = max( 0 , ( mu_s{c}(k) - mu_s{3-c}(k) ) / sqrt( var_s{1}(k) + var_s{2}(k) ) ); I corrected the OV...
- Wed Apr 13, 2016 6:14 pm
- Forum: Box and application development
- Topic: Spectrally weighted CSP OV algorithm box: help to validate ?
- Replies: 6
- Views: 15198
Re: Spectrally weighted CSP OV algorithm box: help to valida
Hi OVibers, The use of the boost multidimensional array library for all the 2d, 3d, 4d (real or complex) data containers http://www.boost.org/doc/libs/1_60_0/libs/multi_array/doc/user.html ... could be a better choice for the algorithm data structures ! (instead of my nested vectors). I am going to ...
- Mon Apr 11, 2016 6:07 pm
- Forum: Box and application development
- Topic: Spectrally weighted CSP OV algorithm box: help to validate ?
- Replies: 6
- Views: 15198
Re: Spectrally weighted CSP OV algorithm box: help to valida
I forgot: the connectivity of the SpecCSPtrainer box is the same than the RegularizedCSPTrainer box: the stimulation signal is the 1st input, the stimulation based signals for class #1 and #2 are the inputs 2 an 3... and it generates a train_completed stimulation signal on the only output (stimulati...
- Mon Apr 11, 2016 4:03 pm
- Forum: Box and application development
- Topic: Spectrally weighted CSP OV algorithm box: help to validate ?
- Replies: 6
- Views: 15198
Spectrally weighted CSP OV algorithm box: help to validate ?
Hi Openvibers, Here is the pre-beta version (well... a code which still needs to be validated functionnally) of the spectrally weighted CSP along the lines of: [1] Tomioka, R., Dornhege, G., Aihara, K., and Mueller, K.-R. "An iterative algorithm for spatio-temporal filter optimization." In Proceedin...
- Wed Feb 17, 2016 4:06 pm
- Forum: Designer
- Topic: Easy way to implement spatiospectral filtering on OV ?
- Replies: 2
- Views: 4290
Re: Easy way to implement spatiospectral filtering on OV ?
Hi Jeff, I'm not sure I totally follow what you want to do, but if your spatiospectral filter is the same thing as a spatiotemporal filter, then if you can represent your spatiotemporal filter as a coefficient matrix, then you could try to hack it as follows 1) Use an epoching box + feature extract...
- Fri Feb 12, 2016 3:11 pm
- Forum: Discussion about OpenViBE
- Topic: Independent Component Analysis on OV
- Replies: 7
- Views: 9879
Re: Independent Component Analysis on OV
Wow, looks super. :) If you'd ever like to contribute these changes back to the upstream, we'd be delighted to include them. Cheers, Jussi Hello Jussi Here are the .h and .cpp files but you might be upset by this little adaptation :shock: :lol: which was just intended to test the Fastica algo with ...
- Sat Feb 06, 2016 7:22 pm
- Forum: Designer
- Topic: Easy way to implement spatiospectral filtering on OV ?
- Replies: 2
- Views: 4290
Easy way to implement spatiospectral filtering on OV ?
Hi Openvibers, Suppose we are able to produce off line some kind of optimal spatiospectral filters (for subsequent classification), is there an easy way to implement several spatiospectral filters on OV ? By spatiospectral filter, I mean several specific pass band filters applied to each channel/sen...
- Sat Feb 06, 2016 6:59 pm
- Forum: Box and application development
- Topic: Sonification box
- Replies: 2
- Views: 10101
Re: Sonification box
Thank you Jussi. It's really a low priority questionning, but It's not as easy as I thought since I would like to be able to listen to the outputs of several spatiospectral filters, transposed / remapped in frequency / filtered (to be audible/compatible with the responses of the varied sound channel...