Holistically-Nested Edge Detection (2015-04-23T00:00:00.000000Z)

TL;DR

HED performs image-to-image prediction by means of a deep learning model that leverages fully convolutional neural networks and deeply-supervised nets, and automatically learns rich hierarchical representations that are important in order to resolve the challenging ambiguity in edge and object boundary detection.

Abstract

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.

Authors

Saining Xie

10 papers

Z. Tu

7 papers

TL;DR

Abstract

Authors

References59 items

Convolutional Oriented Boundaries: From Image Segmentation to High-Level Tasks

Dense Volume-to-Volume Vascular Boundary Detection

Object Skeleton Extraction in Natural Images by Fusing Scale-Associated Deep Side Outputs

Object Contour Detection with a Fully Convolutional Encoder-Decoder Network

A systematic comparison between visual cues for boundary detection

PASCAL Boundaries: A Class-Agnostic Semantic Boundary Dataset

Weakly Supervised Object Boundaries

Pushing the Boundaries of Boundary Detection using Deep Learning

Unsupervised Learning of Edges

Semantic Image Segmentation with Task-Specific Edge Detection Using CNNs and a Discriminatively Trained Domain Transform

Semantic Amodal Segmentation

DeepContour: A deep convolutional feature learned by positive-sharing loss for contour detection

Pixel-wise Deep Learning for Contour Detection

Oriented edge forests for boundary detection

DeepEdge: A multi-scale bifurcated deep network for top-down contour detection

Hypercolumns for object segmentation and fine-grained localization

Fully convolutional networks for semantic segmentation

Deeply-Supervised Nets

Multi-scale Deep Learning for Gesture Detection and Localization

Very Deep Convolutional Networks for Large-Scale Image Recognition

ImageNet Large Scale Visual Recognition Challenge

Learning Rich Features from RGB-D Images for Object Detection and Segmentation

N4-Fields: Neural Network Nearest Neighbor Fields for Image Transforms

The Pascal Visual Object Classes Challenge: A Retrospective

The Role of Context for Object Detection and Semantic Segmentation in the Wild

Fast Edge Detection Using Structured Forests

Visual Boundary Prediction: A Deep Neural Prediction Network and Quality Dissection

Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation

Learning Hierarchical Features for Scene Labeling

Boundary Detection Benchmarking: Beyond F-Measures

Perceptual Organization and Recognition of Indoor Scenes from RGB-D Images

Sketch Tokens: A Learned Mid-level Representation for Contour and Object Detection

Discriminatively Trained Sparse Code Gradients for Contour Detection

Multiscale Convolutional Neural Networks for Vision-Based Classification of Cells

Indoor Segmentation and Support Inference from RGBD Images

Convolutional neural networks applied to house numbers digit classification

Nonparametric Scene Parsing via Label Transfer

Contour Detection and Hierarchical Image Segmentation

Multi-scale Improves Boundary Detection in Natural Images

Auto-context and its application to high-level vision tasks

Recovering Occlusion Boundaries from a Single Image

Supervised Learning of Edges and Object Boundaries

Putting Objects in Perspective

Efficient Graph-Based Image Segmentation

Learning to detect natural image boundaries using local brightness, color, and texture cues

Ecological statistics of Gestalt laws for the perceptual organization of contours.

Neural mechanisms of form and motion processing in the primate visual system

Statistics of Natural Images: Scaling in the Woods

Backpropagation Applied to Handwritten Zip Code Recognition

Scaling Theorems for Zero Crossings

A Computational Approach to Edge Detection

Edge Detection

Scale-space filtering: A new approach to multi-scale description

On the accuracy of the Sobel edge detector

Theory of edge detection

Critical Review : Contour Detection and Hierarchical Image Segmentation

Statistical Edge Detection: Learning and Evaluating Edge Cues

Receptive fields, binocular interaction and functional architecture in the cat's visual cortex

Precision [F=.790] HED (ours) [F=.788] HED−merge (ours) [F=.756] DeepContour [F=.756] CSCNN [F=.753] DeepEdge

Field of Study

Journal Information

Name

Page

Volume

Venue Information

Name

Type

URL

Alternate Names