JOURNAL ARTICLE

Xception-based Stress Classification by Removing ECG Signal’s Outliers

Cheol Ho SongJin-Su KimGyu Ho ChoiSungbum Pan

Year: 2023 Journal:   The Journal of Korean Institute of Information Technology Vol: 21 (5)Pages: 75-81   Publisher: Indonesian Institute of Information Technology

Abstract

스트레스로 인한 질환 발병을 예방하기 위해 심전도 기반의 스트레스 분류 연구가 진행되고 있다. 기존 연구들에서 사용한 디지털 필터링을 통한 방법은 모션 아티팩트(Motion Artifact)와 같이 심전도의 주기적 특성을 훼손하는 잡음을 제거하지 못하는 문제가 발생한다. 그러므로, 본 논문에서는 심전도의 이상치 신호 제거를 이용한 Xception 기반의 스트레스 분류 시스템을 제안한다. 제안하는 시스템은 심전도의 R-R 간격 기반의 유사도 비교를 통해 이상치 신호를 제거한 후 Xception 분류기를 이용한다. 스트레스 분류 실험은 이상치 신호제거 전보다 0.92% 향상된 최대 98.6%의 성능으로, 기존 심전도를 이용한 스트레스 분류 연구들에 비해 2개의 클래스 분류에서 5.6% 이상 향상된 성능을 나타냈다. ECG-based stress classification studies are being conducted to prevent the occurrence of diseases caused by stress. The method through digital filtering used in previous studies has a problem in that it cannot remove noise that damages the periodic characteristics of the ECG, such as motion artifacts. Therefore, in this paper, we propose an Xception-based stress classification system using ECG outlier signal removal. The proposed system uses the Xception classifier after removing the outlier signal through similarity comparison based on the R-R interval of the ECG. The stress classification experiment showed a maximum performance of 98.6%, which is 0.92% higher than before removing the outlier signal, and showed more than 5.6% improved performance in two class classification compared to stress classification studies using ECG.

Keywords:
Outlier Pattern recognition (psychology) Artificial intelligence Classifier (UML) Computer science Artifact (error)

Metrics

2
Cited By
0.67
FWCI (Field Weighted Citation Impact)
13
Refs
0.67
Citation Normalized Percentile
Is in top 1%
Is in top 10%

Citation History

Topics

ECG Monitoring and Analysis
Health Sciences →  Medicine →  Cardiology and Cardiovascular Medicine
Advanced Computing and Algorithms
Social Sciences →  Social Sciences →  Urban Studies
EEG and Brain-Computer Interfaces
Life Sciences →  Neuroscience →  Cognitive Neuroscience

Related Documents

JOURNAL ARTICLE

Removing Outliers

Tom Fearn

Journal:   NIR news Year: 2016 Vol: 27 (5)Pages: 25-25
JOURNAL ARTICLE

ECG Signal’s Classification Algorithm Based on LightGBM

宇光 洪

Journal:   Journal of Image and Signal Processing Year: 2020 Vol: 09 (03)Pages: 165-171
BOOK-CHAPTER

EEG Based Emotion Classification Using Xception Architecture

Arpan PhukanDeepak Gupta

Lecture notes in networks and systems Year: 2022 Pages: 95-108
© 2026 ScienceGate Book Chapters — All rights reserved.