eduzhai > Applied Sciences > Engineering >

MLBF-Net A Multi-Lead-Branch Fusion Network for Multi-Class Arrhythmia Classification Using 12-Lead ECG

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 9 pages

Abstract: Automatic arrhythmia detection using 12-lead electrocardiogram (ECG) signalplays a critical role in early prevention and diagnosis of cardiovasculardiseases. In the previous studies on automatic arrhythmia detection, mostmethods concatenated 12 leads of ECG into a matrix, and then input the matrixto a variety of feature extractors or deep neural networks for extractinguseful information. Under such frameworks, these methods had the ability toextract comprehensive features (known as integrity) of 12-lead ECG since theinformation of each lead interacts with each other during training. However,the diverse lead-specific features (known as diversity) among 12 leads wereneglected, causing inadequate information learning for 12-lead ECG. To maximizethe information learning of multi-lead ECG, the information fusion ofcomprehensive features with integrity and lead-specific features with diversityshould be taken into account. In this paper, we propose a novelMulti-Lead-Branch Fusion Network (MLBF-Net) architecture for arrhythmiaclassification by integrating multi-loss optimization to jointly learningdiversity and integrity of multi-lead ECG. MLBF-Net is composed of threecomponents: 1) multiple lead-specific branches for learning the diversity ofmulti-lead ECG; 2) cross-lead features fusion by concatenating the outputfeature maps of all branches for learning the integrity of multi-lead ECG; 3)multi-loss co-optimization for all the individual branches and the concatenatednetwork. We demonstrate our MLBF-Net on China Physiological Signal Challenge2018 which is an open 12-lead ECG dataset. The experimental results show thatMLBF-Net obtains an average $F 1$ score of 0.855, reaching the highestarrhythmia classification performance. The proposed method provides a promisingsolution for multi-lead ECG analysis from an information fusion perspective.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...