A5014321149- EEG and Brain-Computer Interfaces
- Neural dynamics and brain function
- Blind Source Separation Techniques
- Neural and Behavioral Psychology Studies
- Neuroscience and Neural Engineering
- Photoacoustic and Ultrasonic Imaging
- Sparse and Compressive Sensing Techniques
- Currency Recognition and Detection
- Image and Signal Denoising Methods
- Anomaly Detection Techniques and Applications
- Image Processing Techniques and Applications
- Advanced Memory and Neural Computing
The University of Texas at San Antonio
2012-2017
Broad Institute
2013
KU Leuven
2010
In this paper, we investigated Deep Learning (DL) for characterizing and detecting target images in an image rapid serial visual presentation (RSVP) task based on EEG data. We exploited DL technique with input feature clusters to handle high dimensional features related time - frequency events. The method was applied recordings of a RSVP experiment multiple sessions subjects. For classification non-target images, deep belief net (DBN) classifier the uncorrelated features, which constructed...
The goal of this study is to design an asynchronous steady-state visual evoked potential (SSVEP) BCI system enable control unmanned aerial vehicle (UAV) with multiple commands. An SSVEP-based six different flickering frequencies was constructed realize actuation commands for UAV control. In addition, achieved by including a detection the 'idle' brain state using novel likelihood ratio test and hover command implemented idle state. Offline recording conducted evaluate accuracies game-like...
This paper considers the problem of automatic characterization and detection target images in a rapid serial visual presentation (RSVP) task based on EEG data. A novel method that aims to identify single-trial event-related potentials (ERPs) time-frequency is proposed, robust classifier with feature clustering developed better utilize correlated ERP features. The applied recordings RSVP experiment multiple sessions subjects.The results show image events are mainly characterized by 3 distinct...
Neurotechnologies based on electroencephalography (EEG) and other physiological measures to improve task performance in complex environments will require tools analysis methods that can account for increased environmental noise complexity compared traditional neuroscience laboratory experiments. We propose a bag-of-words (BoW) model address the difficulties associated with realistic applications environments. In this paper, our proof-of-concept results show BoW classifier discriminate two...
Deep learning solutions based on deep neural networks (DNN) and stack (DSN) were investigated for classifying target images in a non-time-locked rapid serial visual presentation (RSVP) image identification task using EEG. Several feature extraction methods associated with this implemented tested learning, where sliding window method the trained classifier was used to predict occurrence of events fashion‥ The algorithms explored stacking able improve error rate by about 5% over existing such...
We consider the detection of control or idle state in an asynchronous Steady-state visually evoked potential (SSVEP)-based brain computer interface system. propose a likelihood ratio test using Canonical Correlation Analysis (CCA) scores calculated from EEG measurements. The exploits state-specific distributions CCA scores. algorithm was tested on offline measurements 42 participants and results should significant improvement error rate over support vector machine classifier. proposed is...
This paper considers the problem of classification electroencephalography (EEG) recordings without precise time locking between stimulus presentation times and recorded EEG waveforms. Traditionally, locking, or perfect timing, information have been crucial in locating region possible neural response. In reality, stimulus' is usually unavailable latency test subjects may not be constant (due to fatigue, concentration, interference, etc.). Therefore, new approaches that do depend on are...
In this paper, the problem of automatic characterizing and detecting target images in an image rapid serial visual presentation (RSVP) task based on EEG data is considered. A novel method that aims at identifying event-related potentials (ERPs) time-frequency proposed, a robust classifier with feature clustering developed to better utilize correlated ERP features. The applied recordings RSVP experiment multiple sessions subjects. results show that, events are mainly characterized by 3...
This work introduces a new technique of robust compression on signals and images, known as Compressive Sensing (CS). It is advanced which can reconstruct sparse from few random acquired samples, achieved to avoid the Nyquist's criteria. Reconstruction signal with just data hard task converted in lineal optimization process various ways find out solution. Widely-used for research at present, CS useful tool sampling/compression that only works signals, moreover we were able implement some kind...