CEFEs: A CNN Explainable Framework for ECG Signals
作者：
Highlights：
• Interpretable explanations by means of functional understanding of internal mechanism of 1D-CNNs for ECG decision support systems, addressing the trust gap found in deep learning.
• Interpretable explanations as metrics that evaluate the quality of CNN models through a set of tri-fold modular tests: a) descriptive statistics, b) visualization and c) detection and mapping of input ECG signal in comparison to its learned feature maps.
• Benchmark for interpretable explanation-based evaluation of CNN models trained for ECG signal classification task.
• Understanding the relationships between machine learned features and the features’ contributions to misclassification of target classes.
• Comparative analysis of machine learned class discriminant features and actual ECG observable diagnostic waveform features.
摘要
•Interpretable explanations by means of functional understanding of internal mechanism of 1D-CNNs for ECG decision support systems, addressing the trust gap found in deep learning.•Interpretable explanations as metrics that evaluate the quality of CNN models through a set of tri-fold modular tests: a) descriptive statistics, b) visualization and c) detection and mapping of input ECG signal in comparison to its learned feature maps.•Benchmark for interpretable explanation-based evaluation of CNN models trained for ECG signal classification task.•Understanding the relationships between machine learned features and the features’ contributions to misclassification of target classes.•Comparative analysis of machine learned class discriminant features and actual ECG observable diagnostic waveform features.
论文关键词：Deep learning,Convolution neural network,ECG Signals,Explainable AI,Explainable Framework,Synthetic healthcare data
论文评审过程：Received 5 July 2020, Revised 18 January 2021, Accepted 23 March 2021, Available online 26 March 2021, Version of Record 5 April 2021.
论文官网地址：https://doi.org/10.1016/j.artmed.2021.102059
