Holistically-Nested Edge Detection

TL;DR

This paper introduces a deep learning-based edge detection method called HED that uses multi-scale, multi-level feature learning and deep supervision to improve accuracy and speed over previous approaches.

Contribution

The paper presents a novel holistic, deeply-supervised CNN architecture for edge detection that advances state-of-the-art performance and efficiency.

Findings

Achieved state-of-the-art F-score of .782 on BSD500

Achieved F-score of .746 on NYU Depth dataset

Significantly faster processing time of 0.4 seconds per image

Abstract

We develop a new edge detection algorithm that tackles 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 approach the human ability resolve the challenging ambiguity in edge and object boundary detection. We significantly advance the state-of-the-art on the BSD500 dataset (ODS F-score of .782) and the NYU Depth dataset (ODS F-score of .746), and do so with an improved speed (0.4 second per image) that is orders of magnitude faster than some recent CNN-based edge detection algorithms.