Semantic Edge Detection with Diverse Deep Supervision

TL;DR

This paper introduces a novel neural network architecture for semantic edge detection that employs diverse deep supervision and an information converter to effectively handle the dual supervision targets of edge localization and semantic categorization.

Contribution

It proposes a multi-task CNN with diverse deep supervision and an information converter to improve semantic edge detection performance.

Findings

Effective on SBD and Cityscapes datasets

Outperforms existing SED methods

Validates the benefit of diverse supervision

Abstract

Semantic edge detection (SED), which aims at jointly extracting edges as well as their category information, has far-reaching applications in domains such as semantic segmentation, object proposal generation, and object recognition. SED naturally requires achieving two distinct supervision targets: locating fine detailed edges and identifying high-level semantics. Our motivation comes from the hypothesis that such distinct targets prevent state-of-the-art SED methods from effectively using deep supervision to improve results. To this end, we propose a novel fully convolutional neural network using diverse deep supervision (DDS) within a multi-task framework where bottom layers aim at generating category-agnostic edges, while top layers are responsible for the detection of category-aware semantic edges. To overcome the hypothesized supervision challenge, a novel information converter unit is introduced, whose effectiveness has been extensively evaluated on SBD and Cityscapes datasets.