ECG Beat Classification on Edge Device

Dennis Hou, Mingzhen Hou, Janpu Hou

2020 · 10 citations · 11 references

Concepts

Abstract

Feature transformation is a class of unsupervised learning methods that has been extensively applied and studied in computer vision. Little work has been done to adapt it to the end-to-end training of visual features on large-scale datasets on edge device. In this work, we present a semi-supervised learning method that jointly learns the parameters of a convolutional neural network (CNN) and using support vector machine (SVM) on the resulting features. By applying the recently developed Saak (Subspace approximation with augmented kernels) Transform, we can update the weights of the network in a single multilayer feed forward step without backpropagation. We apply this semi-supervised training of convolutional neural networks to the MIT-BIH arrhythmia database, on ARM platform such as Raspberry Pi. The resulting model has 98.94% accuracy and 1.06% misclassification rate.

References

11