Hi everyone,
I am relatively new to OpenViBE and I am doing a research with EEG signals for examining the Cognitive Human Properties..I need to perform the following tasks:
1. Visualize the incoming time-based raw data
2. Visualize the previous calculated mental states
3. Calculate the frequency bands
4. Store data sessions in an exportable(CSV) format
5. Build Classifier that recognizes the mental effort.
Out of these tasks, I am able to do 1,3 and 4 tasks until now. But I have problems with calculating mental states and building the classifier for training it. I have attached the scenario file that I designed in the OpenViBE designer and I am using a pre-recorded data for testing it. It would be really helpful if you could suggest me what things I am doing wrong and what I need to improve/try to implement.
Thank you in advance for your time and consideration!
Best Regards!
How to calculated mental states?
-
deepindraw
- Posts: 5
- Joined: Mon May 18, 2015 9:58 am
How to calculated mental states?
- Attachments
-
- PreProcessing and Freq bands.xml
- (145.58 KiB) Downloaded 258 times
-
fabien.lotte
- Posts: 112
- Joined: Sun Mar 14, 2010 12:58 pm
Re: How to calculated mental states?
Dear Deepinraw,
If I understood correctly, your problem is building and using a classifier to recognise a given mental state based on EEG signals, is that right?
If that is so, for your scenario, the problem is that the classifier trainer and the classifier processor are used at the same time (in parallel) in the same scenario. So basically you are using a classifier processor whereas this classifier was not trained yet. Typically, in that case, to do that correctly, we would write one scenario for training the classifier first. And then, you could use it to classify your data in another, second scenario. Also, for the classifier trainer, you need to epoch the data according to which mental state it belongs, to indicate which signals belong to which class (whereas you would not do that when you use the classifier since it is used to predict this class). You should have a look at the motor imagery classification scenarios that come with openvibe to see some examples of how this could be done.
I hope this helps,
Best regards,
Fabien
If I understood correctly, your problem is building and using a classifier to recognise a given mental state based on EEG signals, is that right?
If that is so, for your scenario, the problem is that the classifier trainer and the classifier processor are used at the same time (in parallel) in the same scenario. So basically you are using a classifier processor whereas this classifier was not trained yet. Typically, in that case, to do that correctly, we would write one scenario for training the classifier first. And then, you could use it to classify your data in another, second scenario. Also, for the classifier trainer, you need to epoch the data according to which mental state it belongs, to indicate which signals belong to which class (whereas you would not do that when you use the classifier since it is used to predict this class). You should have a look at the motor imagery classification scenarios that come with openvibe to see some examples of how this could be done.
I hope this helps,
Best regards,
Fabien
-
deepindraw
- Posts: 5
- Joined: Mon May 18, 2015 9:58 am
Re: How to calculated mental states?
Dear Fabien,
Thank you very much for your response. I will try to make different scenarios for classifier trainer and processor and let u know.
Best Regards,
Deepindraw
Thank you very much for your response. I will try to make different scenarios for classifier trainer and processor and let u know.
Best Regards,
Deepindraw