Interpretable arrhythmia diagnosis method in combination with medical field knowledge

Abstract

The invention discloses an electrocardiosignal classification model combining a deep learning model and medical field knowledge, achieving an effect of making compliant and reasonable medical interpretation for classification results while realizing accurate classification of arrhythmia. The method comprises data preprocessing, a deep neural network classification model, a domain knowledge model, a joint training model and an interpretable report model. According to the method, clinical diagnosis rules corresponding to medical pathological features are established, and a deep neural network is combined for joint training. The method has the beneficial effects that 1) for current electrocardiogram diagnosis, only a data driving technology is utilized, medical field knowledge is integrated, and parameters of the neural network are finely adjusted under the guidance of the field knowledge, so that the optimization direction of the deep neural network is more stably related to the field; and 2) an interpretable technology is adopted, visual anomaly positioning is formed through a CAM technology, and in combination with field knowledge, the pathological basis of each diagnosis result is interpreted on the semantic level, so that the medical diagnosis result is more credible.

Description

technical field [0001] The invention relates to the field of biomedical information processing, in particular to a method for diagnosing cardiac arrhythmia that combines deep neural networks and knowledge in the medical field. Background technique [0002] There are 290 million arrhythmia patients in China, and the prevalence rate is increasing every year. The mortality rate of heart disease is 32% higher than that of cancer and other diseases. Early detection and diagnosis are of great significance to reduce mortality and improve quality of life. Electrocardiogram (ECG) is a non-invasive detection method for diagnosing cardiac abnormalities. In recent years, the potential risk of using deep neural networks to detect cardiac abnormalities from ECG signals has been intensively studied, and some projects have achieved better results than humans in specific fields. Better performance by experts. For example, a Stanford University research team led by Andrew Ng used a convoluti...

AI Technical Summary

Problems solved by technology

[0003] At present, there are many wearable products for intelligent diagnosis of ECG signals in the industry, but most of them are only used as daily monitoring tools, and cannot be used as the basis for medical diagnosis. The reason is that in the field of clinical decision-making, the interpretation of neural networks is not enough.

The purely data-driven neural network model can only obtain classification results, but cannot make medical explanations for the results, resulting in unreliable diagnosis results

Images