Classification of normal and abnormal ECG records using lead convolutional neural network and rule inference

With the improvement of living standards, more and more people pay attention to their own health problems and performing an ECG test is the preferred selection for preventing cardiovascular disease. Although it is easy to sample ECG now, diagnostic conclusion cannot be made because of the lack of domain experts, especially in the basic community medical insurance system (BCMIS) of China. A feasible solution is to send ECG records to a cloud-computing platform monitored by ECG physicians, and send back the results to users. In fact, such institutions are common in China now, for instance, Shanghai Aerial Hospital Network. However, there are a large number of ECG records needing to be interpreted and the workload of physicians will be very heavy considering the huge possible audience. For ECG records are mainly collected from people attending physical examinations, their diagnostic conclusion is likely to be “normal”. If a computer-aided ECG diagnostic tool filters out most normal records while physicians only focus on interpreting the remaining abnormal ones, i.e., man-machine integration [1], the diagnostic efficiency will be greatly increased and the social benefits will be significant.