Published Paper Details:

In the contemporary era, cardiovascular diseases, including heart attacks, remain a leading cause of mortality. Identifying heart attack risk in its early stages and accurately assessing such risks related to cardiovascular conditions are crucial for proactive patient management and preventive interventions. This research introduces a pioneering deep learning architecture termed "CardioGraphVision" for the early detection of heart attack risks and cardiovascular conditions using retinal images. The proposed methodology integrates Graph Convolutional Networks (GCNs) and Vision Transformers (ViTs) to incorporate both local structural information and global contextual understanding from retinal scans. A comprehensive simulation analysis is conducted to assess the performance of CardioGraphVision, which is compared against existing algorithms, and its predictive accuracy is measured using pertinent simulation metrics. By conceptualizing retinal images as graph nodes and leveraging self-attention mechanisms, the proposed algorithm achieves precise and efficient feature extraction, essential for early detection of heart attack risk and cardiovascular conditions. To gauge CardioGraphVision's efficacy, diverse simulation experiments are carried out using an extensive dataset of retinal images. Through these comparisons, the predictive accuracy, sensitivity, specificity, and computational efficiency of CardioGraphVision are established. Simulation results demonstrate that CardioGraphVision surpasses existing algorithms in terms of accuracy and sensitivity for early detection of heart attack risk and cardiovascular conditions. Moreover, the algorithm's ability to effectively analyze retinal images with reduced computational overhead further augments its practical relevance.