Nutritional Risk Assessment and Countermeasures for Stroke Patients Based on Deep Learning and Multimodal MRI Images

Abstract

AbstractThe main objective of this study is to evaluate the nutritional risk of stroke patients through deep learning and multimodal MRI images, and provide appropriate nutrition for patients in a timely manner to improve their recovery speed. The method of this article mainly solves the problem of incomplete understanding of patient nutritional risk through a deep learning (DL) and multimodal MRI image-based nutritional risk assessment (RA) model, and accurately provides corresponding countermeasures. The evaluation model based on DL and multimodal MRI images shows that 27 people in Group A are at nutritional risk, accounting for 90%. 26 people in Group B are at nutritional risk, accounting for 86.6%. Both groups of patients urgently need corresponding strategies to reduce risk. Therefore, this article also tested two sets of nutritional support methods, and the results showed that the nutritional support methods in Group B were more effective. The nutritional indicators are not only normal, but the infection rate and mortality rate of patients have also decreased. The results demonstrate that deep learning and multimodal MRI images can promote the recovery process of patients.