eegnet

Deep Learning pipeline for motor-imagery classification.

It is the task to classify BCI competition datasets (EEG signals) using EEGNet and DeepConvNet with different activation functions. You can get some detailed introduction and experimental results in the link below. https://github.com/secondlevel/EEG-classification/blob/main/Experiment%20Report.pdf

The codes that I implemented during my B.Sc. project.

PyTorch code for "Motor Imagery Decoding Using Ensemble Curriculum Learning and Collaborative Training"

Neural control framework for drones using motor imagery EEG classification. Achieves 73% cross-subject accuracy with PyTorch and enables hands-free drone control through imagined hand/feet movements.

Class to automatic create Convolutional Neural Network in PyTorch

NCTU(NYCU) Deep Learning and Practice Spring 2021

Labs for 5003 Deep Learning Practice course in summer term 2021 at NYCU.

This repository contains a replication of the EEGNet code, an efficient convolutional neural network architecture for EEG-based brain–computer interface applications. The project is intended to make the original work more accessible and to serve as a basis for further research and experimentation in EEG signal analysis.

EEG Classification API using Flask

Self-supervised contrastive pretraining for subject-independent motor imagery EEG decoding — fused with Riemannian geometry for few-shot calibration, benchmarked against a classical baseline, with two working Streamlit apps.

Cross-platform EEG signal analysis suite. Features preprocessing pipelines, analysis methods and AI classification for offline research.

Stage training Implementation

EEGNet 脑电解码: 想象左手 vs 想象右手 (PhysioNet EEGMMIDB, PyTorch)

Application that reads real-time EEG signals from a NeuroSky MindWave Mobile 2 headset and uses your brain activity to passively approve or reject code suggestions from Claude Code.

NYCU Deep Learning and Practice Summer 2023

Project for XAI606(Korea University)

Training pipeline for 4-class Motor Imagery (MI) EEG classification on the BCI Competition IV Dataset 2a, using a synthetic / simulated generator that reproduces the event-related desynchronization / synchronization (ERD/ERS) structure described in the project report.

NeuroDecode — EEG Motor Imagery Decoding System (EEGNet/Conformer/TCN/Stacking Ensemble + BrainFlow)

Interpretable EEG-based emotion recognition using EEGNet and Captum (Integrated Gradients), with brain topographic mapping via MNE-Python on the SEED dataset.