Information Technology Reference
In-Depth Information
Fig. 6.
Comparison between results including and not including uncertainty
Table 3.
Oine results introducing uncertainty(%)
Subject 1
Subject 2
Subject 3
123-4
52.1
28.6
19.3
53.1
19.1
27.8
40.3
23.8
35.9
124-3
53.1
14.0
32.9
56.9
11.9
31.2
26.1
24.1
49.8
134-2
67.5
9.5
23.0
55.7
9.6
34.7
25.9
34.2
39.9
234-1
33.4
26.2
40.4
39.6
33.1
22.3
30.9
24.4
44.7
Average
51.5
19.6
28.9
51.3
18.5
30.2
30.8
26.6
42.6
Success
Uncer.
Error Success
Uncer.
Error Success
Uncer.
Error
5Con lu on
A computer vision technique has been proposed to classify mental tasks in a
Brain-Computer Interface. This classifier uses EEG maps obtained from specific
frequencies using data from three different subjects (BCI Competition Data
Set V) and it is based on image comparison. The preliminary results show a
good success percentage of classification so this new classifier can be useful to
differentiate mental tasks in BCI. It is also expected that the results will improve
using visual feedback.
As future works, online tests will be performed using visual feedback to ana-
lyze the usefulness of the classifier.
Search WWH ::
Custom Search