SVCL - Real-Time EEG Surprise Signal Detection

Home People Research Publications Demos

News Jobs Prospective
Students About Internal

Real-Time EEG Surprise Signal Detection

We design a real-time single-trial binary classifier of viewed images based on participant-specific dynamic brain response signatures in high-density (128-channel) electroencephalographic (EEG) data acquired during a rapid serial visual presentation (RSVP) task. Image clips were selected from a broad area image and presented in rapid succession (12/s) in 4.1-s bursts. Participants indicated by subsequent button press whether or not each burst of images included a target airplane feature. Independent component analysis (ICA) was used to extract a set of independent source time-courses and their minimally-redundant low-dimensional informative features in the time and time-frequency amplitude domains from 128- channel EEG data recorded during clip burst presentations in a training session. The cost sensitive boosting algorithm is used to train a high detection rate real-time cascaded classifier for the detection of the target within each frame. The use of Haar features not only allows for computational efficiency leading to real-time performance but also localizes the discriminating features within the EEG data. This can provide insight into the neurological activity that accompany the detection of objects within the human brain. Figure-1 is an example of four of the top ten most discriminating features found by the cost sensitive boosting algorithm along with the component scalp map and the corresponding positive and negative component averages. For example, the top left learned feature automatically exploits post target alpha blocking. The bottom right learned feature exploits eye movement when a target is seen.

Figure-1

Publications: High Detection-rate Cascades for Real-Time Object Detection.
Hamed Masnadi-Shirazi and Nuno Vasconcelos
Proceedings of IEEE International Conference on Computer Vision (ICCV) ,
Rio de Janeiro, Brazil, 2007. � IEEE, [pdf]
Asymmetric Boosting
Hamed Masnadi-Shirazi and Nuno Vasconcelos
Proceedings of International Conference on Machine Learning,
Corvallis, OR, May 2007. [pdf]

Contact: Nuno Vasconcelos, Hamed Masnadi-Shirazi

Home	People	Research	Publications	Demos

News	Jobs	Prospective Students	About	Internal

Real-Time EEG Surprise Signal Detection

We design a real-time single-trial binary classifier of viewed images based on participant-specific dynamic brain response signatures in high-density (128-channel) electroencephalographic (EEG) data acquired during a rapid serial visual presentation (RSVP) task. Image clips were selected from a broad area image and presented in rapid succession (12/s) in 4.1-s bursts. Participants indicated by subsequent button press whether or not each burst of images included a target airplane feature. Independent component analysis (ICA) was used to extract a set of independent source time-courses and their minimally-redundant low-dimensional informative features in the time and time-frequency amplitude domains from 128- channel EEG data recorded during clip burst presentations in a training session. The cost sensitive boosting algorithm is used to train a high detection rate real-time cascaded classifier for the detection of the target within each frame. The use of Haar features not only allows for computational efficiency leading to real-time performance but also localizes the discriminating features within the EEG data. This can provide insight into the neurological activity that accompany the detection of objects within the human brain. Figure-1 is an example of four of the top ten most discriminating features found by the cost sensitive boosting algorithm along with the component scalp map and the corresponding positive and negative component averages. For example, the top left learned feature automatically exploits post target alpha blocking. The bottom right learned feature exploits eye movement when a target is seen.

Figure-1
Publications:	High Detection-rate Cascades for Real-Time Object Detection. Hamed Masnadi-Shirazi and Nuno Vasconcelos Proceedings of IEEE International Conference on Computer Vision (ICCV) , Rio de Janeiro, Brazil, 2007. � IEEE, [pdf]
	Asymmetric Boosting Hamed Masnadi-Shirazi and Nuno Vasconcelos Proceedings of International Conference on Machine Learning, Corvallis, OR, May 2007. [pdf]
Contact:	Nuno Vasconcelos, Hamed Masnadi-Shirazi