AI-Driven Early Detection of Cardiovascular Disorders Using Multimodal Imaging Data

Abstract

The current paper introduces an AI-based hybrid deep learning solution to the task of detecting cardiovascular conditions in their early stages, with multimodal imaging data. The suggested approach combines the Convolutional Neural Networks (CNN), the Transformer Networks, and the Long Short-Term Memory (LSTM) to analyze the image and time data of a complex format, enhancing the strength and precision of the cardiovascular disease forecasts. The use of Federated Learning is to guarantee the privacy of data and to allow medical institutions to collaborate in training models. Furthermore, to improve the interpretability of the model, the Attention Mechanisms are also introduced to help the model concentrate on the most relevant items in both the imaging and the patient history data. The model is implemented on the Microsoft Azure machine learning, which offers scalability and computing capabilities that are required to efficiently train and make predictions in real-time. The results of the experiments show that the given method has an accuracy of 92, which is higher than the traditional methods. The research presents a privacy-preserving, scalable, and efficient framework of detecting cardiovascular disorders in their early stages, and it can be applied in a clinical environment.

Ähnliche Arbeiten