Accelerating Image Classification with Graph Convolutional Neural Networks using Voronoi Diagrams

--Recent advances in image classification have been significantly propelled by the integration of Graph Convolutional Networks (GCNs), offering a novel paradigm for handling complex data structures. This study introduces an innovative framework that employs GCNs in conjunction with V oronoi diagrams to peform image classification, leveraging their exceptional capability to model relational data. Unlike conventional convolutional neural networks, our approach utilizes a graph-based representation of images, where pixels or regions are treated as vertices of a graph, which are then simplified in the form of the corresponding Delaunay triangulations. Our model yields significant improvement in pre-processing time and classification accuracy on several benchmark datasets, surpassing existing state-of-the-art models, especially in scenarios that involve complex scenes and fine-grained categories. The experimental results, validated via cross-validation, underscore the potential of integrating GCNs with V oronoi diagrams in advancing image classification tasks. This research contributes to the field by introducing a novel approach to image classification, while opening new avenues for developing graph-based learning paradigms in other domains of computer vision and non-structured data. In particular, we have proposed a new version of the GCN in this paper, namely normalized V oronoi Graph Convolution Network (NVGCN), which is faster than the regular GCN. The domain of image classification has witnessed a paradigm shift with the advent of deep learning techniques, particularly Graph Convolutional Neural Networks (GCNs), which have revolutionized how we approach complex tasks on image data. GCNs, by their nature, are adept at handling data represented in a graph format, making them an ideal choice for tasks where relational context and structural information are pivotal. This introduction outlines recent advancements in image classification using GCNs, focusing on the use of superpixels and wavelet techniques. GCNs have emerged as a powerful paradigm in image classification, primarily due to their ability to capture and process the non-Euclidean structure of the data. Zhou, et al. provided a comprehensive overview of the applications of GCNs in various fields, emphasizing their efficacy in image classification tasks [1].