计算机科学
卷积神经网络
人工智能
边缘计算
短时傅里叶变换
特征提取
模式识别(心理学)
推论
边缘设备
深度学习
特征(语言学)
机器学习
GSM演进的增强数据速率
云计算
傅里叶变换
数学
傅里叶分析
语言学
哲学
数学分析
操作系统
出处
期刊:IEEE Access
[Institute of Electrical and Electronics Engineers]
日期:2022-01-01
卷期号:10: 94469-94486
标识
DOI:10.1109/access.2022.3204703
摘要
Automated classification of Electrocardiogram (ECG) for arrhythmia monitoring is the core of cardiovascular disease diagnosis. Machine Learning (ML) is widely used for arrhythmia detection using various feature engineering and classification models. Cloud-based inference is the prevailing deployment model of modern ML algorithms which does not always meet the availability and privacy requirements of ECG monitoring. Edge inference is an emerging alternative which addresses the concerns of latency, privacy, connectivity, and availability. However, edge deployment of AI models is challenging due to the demanding requirements of modern ML algorithms and the computation constraints of edge devices. In this work, we propose a lightweight self-contained short-time Fourier Transform (STFT) Convolutional Neural Network (CNN) model for ECG classification and arrhythmia detection in real-time at the edge. We provide a clear interpretation of the convolutional layer as a Finite Impulse Response (FIR) filter and exploit this interpretation to develop an STFT-based 1D convolutional (Conv1D) layer to extract the spectrogram of the input ECG signal. The Conv1D output feature maps are reshaped into a 2D heatmap and fed to a 2D convolutional (Conv2D) neural network (CNN) for classification. The real-time performance of the proposed model is planned in advance to fit the resource constraints of edge inference. Four model variants are trained, optimized, and tested on a raspberry-pi edge device. Weight quantization and pruning techniques are applied to the developed models to optimize them for edge computing. The MIT-BIH arrhythmia database is used for model training and testing. The proposed classifiers can achieve up to 99.1% classification accuracy and 95% F1-score at the edge with a maximum model size of 90 KB, an average inference time of 9 ms, and a maximum memory usage of 12 MB. The achieved results of the proposed classifier enable its deployment on a wide range of edge devices for arrhythmia detection.
科研通智能强力驱动
Strongly Powered by AbleSci AI