Practical fine-grained learning based anomaly classification for ECG image

作者：

Highlights：

• Performance of the existing signal based ECG algorithm declines when applying to ECG images.

• The ecg data stored in image format has unique characteristics that distinguish it from the time digital format.

• Fine-grained based algorithms can effectively identify subtle differences between different types of ECG abnormalities.

摘要

•Performance of the existing signal based ECG algorithm declines when applying to ECG images.•The ecg data stored in image format has unique characteristics that distinguish it from the time digital format.•Fine-grained based algorithms can effectively identify subtle differences between different types of ECG abnormalities.

论文关键词：Neural networks,ECG anomaly classification,Machine learning,Fine-grained classification

论文评审过程：Received 30 September 2020, Revised 27 May 2021, Accepted 22 June 2021, Available online 21 July 2021, Version of Record 25 August 2021.

论文官网地址：https://doi.org/10.1016/j.artmed.2021.102130