Towards Precision Cardiac Healthcare: Deep Learning and IoT Integration for Real-Time Monitoring and Personalized Diagnosis

The growing prevalence of cardiovascular diseases (CVDs) underscores the critical need for accurate, real-time monitoring and personalized diagnostics in cardiac healthcare. This paper presents DeepCardioNet, an innovative deep learning framework integrated with Internet of Things (IoT) devices, designed to address the limitations of existing cardiac health monitoring systems. The proposed framework leverages multimodal physiological data, including electrocardiogram (ECG), heart rate variability (HRV), and blood pressure (BP), to provide comprehensive and precise diagnostics. DeepCardioNet incorporates a multi-stream convolutional neural network and transformer architecture, enhanced with an attention mechanism for dynamic feature prioritization. In addition, we optimize the model through techniques such as pruning and quantization to support the possibility of real-time edge AI deployment, ensuring efficient inference on resource-constrained IoT devices. Experimental results on the MIT-BIH arrhythmia database and MIMIC-III waveform database demonstrate the superiority of DeepCardioNet over state-of-the-art methods, achieving an accuracy of 93.6% and an F1-score of 91.8%. Extensive ablation studies validate the contributions of key architectural components, including multimodal data fusion and attention mechanisms. The findings highlight the framework’s robustness and suitability for real-world cardiac health monitoring applications. By integrating advanced AI techniques with IoT capabilities, DeepCardioNet represents a significant step toward achieving personalized and precise healthcare in the context of the healthcare industry.