Spatial Pyramid Based Graph Reasoning for Semantic Segmentation
This work addresses the problem of global modeling in semantic segmentation for computer vision applications, offering an incremental improvement with a lightweight module.
The paper tackles the limited receptive field in convolution for semantic segmentation by proposing a data-dependent Laplacian with an attention diagonal matrix for graph reasoning in the original feature space, achieving comparable performance with reduced computational and memory overhead on datasets like Cityscapes and COCO Stuff.
The convolution operation suffers from a limited receptive filed, while global modeling is fundamental to dense prediction tasks, such as semantic segmentation. In this paper, we apply graph convolution into the semantic segmentation task and propose an improved Laplacian. The graph reasoning is directly performed in the original feature space organized as a spatial pyramid. Different from existing methods, our Laplacian is data-dependent and we introduce an attention diagonal matrix to learn a better distance metric. It gets rid of projecting and re-projecting processes, which makes our proposed method a light-weight module that can be easily plugged into current computer vision architectures. More importantly, performing graph reasoning directly in the feature space retains spatial relationships and makes spatial pyramid possible to explore multiple long-range contextual patterns from different scales. Experiments on Cityscapes, COCO Stuff, PASCAL Context and PASCAL VOC demonstrate the effectiveness of our proposed methods on semantic segmentation. We achieve comparable performance with advantages in computational and memory overhead.