Heterogeneous Grid Convolution for Adaptive, Efficient, and Controllable Computation
Heterogeneous Grid Convolution for Adaptive, Efficient, and Controllable Computation
复制标题
DOI：
10.1109/cvpr46437.2021.01373
复制
发表时间：
2021-04
期刊：
影响因子：
--
通讯作者：
Ryuhei Hamaguchi;Yasutaka Furukawa;M. Onishi;Ken Sakurada
中科院分区：
文献类型：
--
作者：
Ryuhei Hamaguchi;Yasutaka Furukawa;M. Onishi;Ken Sakurada
本文提出了一种新颖的异构网格卷积，它通过利用图像内容的异构性来构建基于图的图像表示，从而在卷积架构中实现自适应、高效和可控的计算。更具体地说，该方法通过可微分聚类方法从卷积层构建数据自适应图结构，将特征池化到图，执行新颖的方向感知图卷积，并将特征反池化回卷积层。通过使用开发的模块，本文提出了异构网格卷积网络，高效且对现有架构的强大扩展。我们在四个图像理解任务、语义分割、对象定位、道路提取和显着对象检测上评估了所提出的方法。所提出的方法对四项任务中的三项有效。特别是，该方法优于强大的基线，语义分割的浮点运算减少了 90% 以上，并实现了道路提取的最先进结果。我们将分享我们的代码、模型和数据。
This paper proposes a novel heterogeneous grid convolution that builds a graph-based image representation by exploiting heterogeneity in the image content, enabling adaptive, efficient, and controllable computations in a convolutional architecture. More concretely, the approach builds a data-adaptive graph structure from a convolutional layer by a differentiable clustering method, pools features to the graph, performs a novel direction-aware graph convolution, and unpool features back to the convolutional layer. By using the developed module, the paper proposes heterogeneous grid convolutional networks, highly efficient yet strong extension of existing architectures. We have evaluated the proposed approach on four image understanding tasks, semantic segmentation, object localization, road extraction, and salient object detection. The proposed method is effective on three of the four tasks. Especially, the method outperforms a strong baseline with more than 90% reduction in floating-point operations for semantic segmentation, and achieves the state-of-the-art result for road extraction. We will share our code, model, and data.