2021 KNU 전자공학설계 팀프로젝트

Artigence - 딥러닝을 활용한 BCI기반 기술 응용

Members

Name	github_ID
박재성	wotjd0715
김태헌	TaeHeonKim250
장현진	nowwhy
신예주
김민채
김예지
김지혜	jihyekim213

Code

Deep_learning_model 참고

Cross_Newmodel.py : Cross-subject performance with proposed Multi-DS-EEGNet
Cross_Subject.py : Cross-subject performance with EEGNet
NewModel.py : Within-subject performance with proposed Multi-DS-EEGNet
Within_Subject.py : Within-subject performance with EEGNet
EEGNet.py : Several EEGNet version with proposed Multi-DS-EEGNet
preprocessing.m : Matlab code for preprocessing(BPF,ASR,CAR,ICA) with eeglab

Tech

공개 MI dataset을 이용하여 Python 및 Tensorflow를 통해 EEGNET을 학습시켜 BMI 설계 및 제작
Open DB : http://deepbci.korea.ac.kr/opensource/opendb/

Progress ✨

Summary 참고

Our Workplan

주제

2학기

1학기

Project 진행상황

2학기

Idea meeting picture

DataSet we used

---

BCI System process

---

Signal Acquisition

---

Signal Processing

classification

Classification Algorithm

---

Application

Prototype

구상도

Flow Diagram

---

Preprocessing

How to Preprocess Dataset

BPF, ASR, CAR, ICA in Matlab

Preprocessing

BPF(Band Pass Filter)

50Hz이상 주파수 제거

ASR(Artifact Subspace Reconstruction)

bad channel 제거

CAR(Common Average Reference)

전위 기준점 변경

ICA(Independent Component Analysis)

Noise가 될 수 있는 요인 제거 ex) Eye Blinking, Heart Rate...