All Projects
2022
BCI Signal Classification
Classification models for Symmetric Positive Definite (SPD) matrices using Riemannian geometry methods, applied to EEG brain-signal data from Brain-Computer Interface systems at Auburn University.
SPD matrix classification
BCI systems
Riemannian geometry
EEG data analysis
Python
scikit-learn
NumPy
Riemannian Geometry
EEG
Statistics
Overview
Research project at Auburn University developing and evaluating classification algorithms for Brain-Computer Interface (BCI) systems. Applied Riemannian geometry methods to classify Symmetric Positive Definite (SPD) matrices derived from electroencephalogram (EEG) brain-signal data.
Methods
- Implemented classification models operating on the Riemannian manifold of SPD matrices
- Evaluated algorithms on EEG data from standard BCI benchmark datasets
- Compared convergence rate, computational complexity, and classification accuracy across methods
Results
- Demonstrated improved classification accuracy using Riemannian geometry vs. Euclidean baselines
- Comprehensive algorithm comparison across multiple EEG paradigms
- Findings contributed to Auburn University research on BCI signal processing