Enhanced Structural Image Classification using Hybrid CNN-GNN Model

The availability of structural image data has significantly increased in recent years due to the widespread use of improved imaging techniques. The increase in the amount of data has required the creation of strong and effective classification algorithms to extract valuable information from these photos. Convolutional Neural Networks (CNNs) have exhibited exceptional efficacy in image classification tasks by capturing hierarchical characteristics. Graph Neural Networks (GNNs) have demonstrated potential in representing intricate connections within graph-structured data. This research presents a new hybrid model that combines the advantages of both CNNs and GNNs to improve the categorization of structural images. Our model combines the spatial locality acquired by CNNs with the relational information stored by GNNs to obtain better classification performance. We perform extensive experiments on standard structural picture datasets to showcase the efficacy of our suggested methodology. The findings demonstrate that our hybrid CNN-GNN model surpasses current approaches, underscoring its capacity to enhance structural picture categorization tasks.