Strip Pooling: Rethinking Spatial Pooling for Scene Parsing
This work addresses the challenge of improving spatial pooling for scene parsing, offering a lightweight, plug-and-play module that enhances performance in pixel-wise prediction tasks, though it is incremental as it builds on existing pooling techniques.
The paper tackles the problem of capturing long-range contextual information in scene parsing by introducing strip pooling, a novel spatial pooling strategy using long, narrow kernels, and demonstrates that this approach achieves new state-of-the-art results on benchmarks like ADE20K and Cityscapes.
Spatial pooling has been proven highly effective in capturing long-range contextual information for pixel-wise prediction tasks, such as scene parsing. In this paper, beyond conventional spatial pooling that usually has a regular shape of NxN, we rethink the formulation of spatial pooling by introducing a new pooling strategy, called strip pooling, which considers a long but narrow kernel, i.e., 1xN or Nx1. Based on strip pooling, we further investigate spatial pooling architecture design by 1) introducing a new strip pooling module that enables backbone networks to efficiently model long-range dependencies, 2) presenting a novel building block with diverse spatial pooling as a core, and 3) systematically comparing the performance of the proposed strip pooling and conventional spatial pooling techniques. Both novel pooling-based designs are lightweight and can serve as an efficient plug-and-play module in existing scene parsing networks. Extensive experiments on popular benchmarks (e.g., ADE20K and Cityscapes) demonstrate that our simple approach establishes new state-of-the-art results. Code is made available at https://github.com/Andrew-Qibin/SPNet.