We present Convolutional Oriented Boundaries (COB), which produces multiscale oriented contours and region hierarchies starting from generic image classification Convolutional Neural Networks (CNNs). COB is computationally efficient, because it requires a single CNN forward pass for multi-scale contour detection and it uses a novel sparse boundary representation for hierarchical segmentation; it gives a significant leap in performance over the state-of-the-art, and it generalizes very well to unseen categories and datasets. Particularly, we show that learning to estimate not only contour strength but also orientation provides more accurate results. We perform extensive experiments for low-level applications on BSDS, PASCAL Context, PASCAL Segmentation, and NYUD to evaluate boundary detection performance, showing that COB provides state-of-the-art contours and region hierarchies in all datasets. We also evaluate COB on high-level tasks when coupled with multiple pipelines for object proposals, semantic contours, semantic segmentation, and object detection on various databases (MS-COCO, SBD, PASCAL VOC'07), showing that COB also improves the results for all tasks.

本文提出了一种名为Convolutional Oriented Boundaries (COB)的方法，该方法通过图像分类卷积神经网络（CNNs）产生多尺度的定向轮廓和区域层次结构，其具有高效的计算性能和显著的识别性能，并且泛化性能也很好。我们在多个数据集上验证了COB的性能，并通过与不同任务的结合，如目标提议，语义轮廓，语义分割和目标检测等，证明COB在提高多种任务的性能方面也具有良好的效果。

卷积定向边界: 从图像分割到高级任务