Impact Statement:Grounded in cognitive science, brain science, and neuroscience, relational reasoning and structured modeling are an integral part of human cognition. Geometric representa...Show More
Abstract:
A graph structure is a powerful mathematical abstraction, which can not only represent information about individuals but also capture the interactions between individuals...Show MoreMetadata
Impact Statement:
Grounded in cognitive science, brain science, and neuroscience, relational reasoning and structured modeling are an integral part of human cognition. Geometric representation and relational modeling of image content is an important and challenging task for advanced intelligent perception and image interpretation. Graph representation learning algorithms provide an essential way to address this problem. This paper systematically reviews graph representation learning methods and its application for visual tasks. The benefits of this paper are, firstly, this is the first comprehensive overview of graph representations in computer vision community. Second, it provides a theoretical background and practical guidelines for researchers who attempt to improve the robustness of image processing tasks with graph representations. Third, with the biological and brain-inspired context, this paper discusses open issues and future directions for graph representation and computer vision.
Abstract:
A graph structure is a powerful mathematical abstraction, which can not only represent information about individuals but also capture the interactions between individuals for reasoning. Geometric modeling and relational inference based on graph data is a long-standing topic of interest in the computer vision community. In this article, we provide a systematic review of graph representation learning and its applications in computer vision. First, we sort out the evolution of representation learning on graphs, categorizing them into the nonneural network and neural network methods based on the way the nodes are encoded. Specifically, nonneural network methods, such as graph embedding and probabilistic graphical models, are introduced, and neural network methods, such as graph recurrent neural networks, graph convolutional networks, and variants of graph neural networks, are also presented. Then, we organize the applications of graph representation algorithms in various vision tasks (such...
Published in: IEEE Transactions on Artificial Intelligence ( Volume: 4, Issue: 1, February 2023)