CASENet: Deep Category-Aware Semantic Edge Detection

TL;DR

This paper introduces CASENet, a deep learning architecture for category-aware semantic edge detection that models edges as multi-label problems, significantly outperforming previous methods on standard datasets.

Contribution

The paper presents a novel end-to-end deep architecture with a multi-label loss for semantic edge detection, handling multi-class edges more effectively than prior approaches.

Findings

Outperforms state-of-the-art on SBD and Cityscapes datasets

Uses a multi-label loss for better edge classification

Employs a skip-layer architecture for improved feature fusion

Abstract

Boundary and edge cues are highly beneficial in improving a wide variety of vision tasks such as semantic segmentation, object recognition, stereo, and object proposal generation. Recently, the problem of edge detection has been revisited and significant progress has been made with deep learning. While classical edge detection is a challenging binary problem in itself, the category-aware semantic edge detection by nature is an even more challenging multi-label problem. We model the problem such that each edge pixel can be associated with more than one class as they appear in contours or junctions belonging to two or more semantic classes. To this end, we propose a novel end-to-end deep semantic edge learning architecture based on ResNet and a new skip-layer architecture where category-wise edge activations at the top convolution layer share and are fused with the same set of bottom layer features. We then propose a multi-label loss function to supervise the fused activations. We show that our proposed architecture benefits this problem with better performance, and we outperform the current state-of-the-art semantic edge detection methods by a large margin on standard data sets such as SBD and Cityscapes.