⭐⭐⭐⭐⭐ IX Jornadas Académicas y I Congreso Científico de Ciencias e Ingeniería: Clasificación de señales de Electroencefalografía #EEG con Redes Neuronales #NN en #FPGA
0 likes8,222 views
The document discusses the classification of electroencephalography (EEG) signals using neural networks on FPGA technology, detailing the methodology involving preprocessing, feature extraction, and classification of different brain wave frequencies. It presents results from an experiment with 64 electrodes measuring motor tasks of hands and feet, highlighting significant findings in the motor cortex. Future work includes developing robotic arm control using EEG signals and mentions achievements in related international competitions.
![Analysis of Results
Clusters:
1. T3 Motor activity/tasks of both hands
2. T4 Motor activity/tasks of both feet
Time to open the file in the SD 21,26 [us]
Time to open the file in the SD 22,30 [us]
Processing time of the neural
network
27,36 [us]](https://image.slidesharecdn.com/unemi-210228212034/85/IX-Jornadas-Academicas-y-I-Congreso-Cientifico-de-Ciencias-e-Ingenieria-Clasificacion-de-senales-de-Electroencefalografia-EEG-con-Redes-Neuronales-NN-en-FPGA-10-320.jpg)
More Related Content
Similar to ⭐⭐⭐⭐⭐ IX Jornadas Académicas y I Congreso Científico de Ciencias e Ingeniería: Clasificación de señales de Electroencefalografía #EEG con Redes Neuronales #NN en #FPGA (20)
⭐⭐⭐⭐⭐ IX Jornadas Académicas y I Congreso Científico de Ciencias e Ingeniería: Clasificación de señales de Electroencefalografía #EEG con Redes Neuronales #NN en #FPGA
- 1. Clasificación de señales de Electroencefalografía (EEG) con redes neuronales en FPGA Víctor Asanza
- 2. FPGA (Field-Programmable Gate Array) Arreglos de Puertas Programables por Campos Aplicaciones con FPGAs
- 3. FPGA DE10-STANDARD System on Programmable Chip (SoPC)
- 4. EEG SIGNALS 64 surface EEG Electrodes International System 10-20 DC artifact present on the 64 electrodes of the EEG signal
- 5. 0.5–4 Hz • Delta waves • Sleep REM 4 –8 Hz • Theta waves • Meditation 8 –15 Hz • Alpha waves • Relax • μ waves (8-12Hz) • Imaginary Motor 16-31 Hz • Beta waves • Alert 32 –110 Hz • Gamma waves EEG SIGNALS
- 6. Data Set • Emotiv Epoc • 25 Healthy subjects • Sampling frequency of 160Hz • Task (open and close both hands or both feet) • Motor activity/tasks of both hands (E1) • Motor activity/tasks of both feet (E2)
- 7. Methodology and Results EEG Signals Signals Preprocessing Features Extraction Features Selection Classification Frequency analysis with the FFT of the original EEG signals Bandpass filter Buttherworth-IIR, 7-30 Hz Frequency analysis with the FFT of the filtered EEG signals
- 8. Methodology and Results EEG Signals Signals Preprocessing Features Extraction Features Selection Classification • A periodogram (Welch PSD) • Power Spectral Density (PSD) features • Maximum PSD value • Frequency • Arithmetic mean • Variance • 64 electrodes x 4 features
- 9. Methodology and Results EEG Signals Signals Preprocessing Features Extraction Features Selection Classification 21 x 4 features in the imaginary motor task both hands Maximum PSD value and frequency occur in the 21 electrodes located in the motor cortex
- 10. Analysis of Results Clusters: 1. T3 Motor activity/tasks of both hands 2. T4 Motor activity/tasks of both feet Time to open the file in the SD 21,26 [us] Time to open the file in the SD 22,30 [us] Processing time of the neural network 27,36 [us]
- 11. Trabajos en curso Control de brazo robótico usando redes neuronales en señales EEG
- 12. Resultados Obtenidos Paper aceptado en conferencia internacional “k-NN-based EMG recognition for gestures communication with limited hardware resources” http://www.smart-world.org/2019/uic/
- 13. Resultados Obtenidos Tercer Lugar en concurso internacional “Artificial Neural Network based EMG recognition for gesture communication” http://www.innovatefpga.com/cgi-bin/innovate/teams.pl?Id=AS027
- 15. Víctor Asanza Mail: [email protected] Facultad de Ingeniería en Electricidad y Computación, FIEC Escuela Superior Politécnica del Litoral, ESPOL Campus Gustavo Galindo Km 30.5 Vía Perimetral, P.O. Box 09-01-5863 090150 Guayaquil, Ecuador For more information
- 16. Thank you!