CT Image Data Analysis Framework Based on Computer Deep Learning: Cardiovascular Disease Prediction

Cardiovascular disease affects people's healthy life, but traditional prediction methods that rely on limited clinical indicators are difficult to fully capture the complexity and individual differences of the disease. In this study, CNN-LSTM is used to deeply mine and analyze CT image data in order to accurately predict the risk of cardiovascular disease. In this paper, a large number of CT image data including patients and non-patients with cardiovascular diseases are collected, and the combination of Convolutional Neural Network (CNN) and Long-term Memory Network (LSTM) is used to extract features and analyze time series of CT images. CNN is responsible for extracting spatial features from images, while LSTM is used to capture the time dependence in image sequences. Firstly, this paper preprocesses the data, designs and trains a model that combines CNN and LSTM. In the process of model training, the transfer learning strategy is adopted, and then the pre-trained CNN model is used to extract features and optimize the model parameters. CNN-LSTM model performs well in cardiovascular disease prediction, with the prediction accuracy of disease types as high as 97.65%, the prediction accuracy of recurrence rate between 95.3% and 99.9%, and the prediction accuracy of severity remains above 92%, which is better than XGBoost model. This paper provides a comprehensive perspective for cardiovascular risk assessment, assists clinical diagnosis, and has high clinical application potential.