Holistically-Nested Edge Detection
Holistically-Nested Edge Detection
复制标题
DOI：
10.1007/s11263-017-1004-z
复制
发表时间：
2017-12-01
影响因子：
19.5
通讯作者：
Tu, Zhuowen
中科院分区：
文献类型：
--
作者：
Xie, Saining;Tu, Zhuowen
我们开发了一种新的边缘检测算法，该算法解决了这个长期视觉问题中的两个重要问题：（1）整体图像训练和预测； （2）多尺度和多级特征学习。我们提出的方法是整体上巢的边缘检测（HED），通过一种深度学习模型来实现图像到图像的预测，该模型利用完全卷积的神经网络和深度监督的网络。 HED会自动学习丰富的层次结构表示（以深度监督为指导），这些表示对于解决边缘和对象边界检测的挑战性歧义很重要。我们在BSDS500数据集（ODS F-评分为0.790）和NYU深度数据集（ODS F-SCORE 0.746）上大大提高了最先进的方法，并以提高的速度（每图像0.4 s）这样做这是比HED之前开发的一些基于CNN的边缘检测算法快的数量级。我们还观察到其他边界检测基准数据集（如多核和Pascal-Context）的令人鼓舞的结果。
We develop a new edge detection algorithm that addresses two important issues in this long-standing vision problem: (1) holistic image training and prediction; and (2) multi-scale and multi-level feature learning. Our proposed method, holistically-nested edge detection (HED), performs image-to-image prediction by means of a deep learning model that leverages fully convolutional neural networks and deeply-supervised nets. HED automatically learns rich hierarchical representations (guided by deep supervision on side responses) that are important in order to resolve the challenging ambiguity in edge and object boundary detection. We significantly advance the state-of-the-art on the BSDS500 dataset (ODS F-score of 0.790) and the NYU Depth dataset (ODS F-score of 0.746), and do so with an improved speed (0.4 s per image) that is orders of magnitude faster than some CNN-based edge detection algorithms developed before HED. We also observe encouraging results on other boundary detection benchmark datasets such as Multicue and PASCAL-Context.