Heart Image Digital Model Building and Feature Extraction Analysis Based on Deep Learning

Abstract

Heart image digital model building and feature extraction analysis based on deep learning is proposed in this research. In the cardiac MRI, the blood of high-speed movement often has artifacts, gray is uneven, causing the blood to the heart muscle contrast between the lower, inside and outside of the left ventricle contour segmentation difficult. In general, in the course of the diastolic image quality is poor; only during the diastolic and heart image quality is the best. Because you left the room filled with blood, the heart is in a relatively static state. Therefore, we use the digital system with deep learning to analyze the extracted images. Due to the threshold segmentation of the original heart image directly. The effect is very poor. The foreground and background cannot be discerned, but the threshold segmentation of the heart image is based on morphological reconstruction. Get a rough area of the heart and based on this, we carry out foreground, background mark. Experiments using a disc structure to perform morphological erosion on the above areas to obtain foreground marks. The simulation results prove the effectiveness of the model. The segmentation accuracy is higher than the other methodologies.