Cardiovascular Disease Diagnosis Method Based on AMEsim Simulation and Convolutional Neural Networks

Traditional cardiovascular disease diagnosis heavily relies on the doctor’s experience. This paper proposes a diagnostic method for cardiovascular diseases that combines the AMEsim software with a Convolutional Neural Network (CNN) model. First, an AMEsim simulation model of the cardiovascular system was established and verified through theoretical calculations and clinical experiments. Then, various faults were introduced into the model to simulate different pathological conditions, and data were extracted to build a dataset. Finally, a CNN model was developed, trained using the dataset, and optimized with a genetic algorithm to extract features from different pathological conditions for diagnosis. The results show that the CNN model, trained using multi-channel data and optimized by the genetic algorithm, outperforms traditional single-channel, dual-channel, and multi-channel models in diagnosing different pathological conditions simulated by the model. The genetic algorithm-optimized multi-channel model achieved a comprehensive diagnostic accuracy of 100% on the test set, representing a significant improvement over the single-channel model and the dual-channel model in most cases. Moreover, the number of iterations required for training was only seven, surpassing all traditional single-channel, dual-channel, and multi-channel models, resulting in lower training time costs. This provides an efficient and accurate new method for diagnosing cardiovascular diseases, offering diagnostic insights for clinical practice.