Dynamic Graph Message Passing Networks
This work addresses the computational inefficiency of fully-connected graphs for long-range dependency modelling in computer vision, offering a more efficient solution for scene understanding tasks.
The paper tackles the problem of capturing long-range dependencies in scene understanding tasks by proposing a dynamic graph message passing network that adaptively samples nodes and predicts node-dependent filter weights, achieving significant improvements over state-of-the-art baselines on three tasks with fewer floating-point operations and parameters.
Modelling long-range dependencies is critical for scene understanding tasks in computer vision. Although CNNs have excelled in many vision tasks, they are still limited in capturing long-range structured relationships as they typically consist of layers of local kernels. A fully-connected graph is beneficial for such modelling, however, its computational overhead is prohibitive. We propose a dynamic graph message passing network, that significantly reduces the computational complexity compared to related works modelling a fully-connected graph. This is achieved by adaptively sampling nodes in the graph, conditioned on the input, for message passing. Based on the sampled nodes, we dynamically predict node-dependent filter weights and the affinity matrix for propagating information between them. Using this model, we show significant improvements with respect to strong, state-of-the-art baselines on three different tasks and backbone architectures. Our approach also outperforms fully-connected graphs while using substantially fewer floating-point operations and parameters. The project website is http://www.robots.ox.ac.uk/~lz/dgmn/