End-to-End Instance Edge Detection

TL;DR

This paper introduces a novel transformer-based method for precise instance edge detection that leverages high-resolution features and point supervision, achieving competitive results and complementing existing segmentation and detection tasks.

Contribution

The paper proposes a new transformer architecture with a lightweight dense prediction head and a penalty reduced focal loss for efficient, high-resolution instance edge detection.

Findings

Achieves state-of-the-art performance in instance edge detection.

Effective point supervision reduces annotation costs.

Complementary to existing segmentation and detection methods.

Abstract

Edge detection has long been an important problem in the field of computer vision. Previous works have explored category-agnostic or category-aware edge detection. In this paper, we explore edge detection in the context of object instances. Although object boundaries could be easily derived from segmentation masks, in practice, instance segmentation models are trained to maximize IoU to the ground-truth mask, which means that segmentation boundaries are not enforced to precisely align with ground-truth edge boundaries. Thus, the task of instance edge detection itself is different and critical. Since precise edge detection requires high resolution feature maps, we design a novel transformer architecture that efficiently combines a FPN and a transformer decoder to enable cross attention on multi-scale high resolution feature maps within a reasonable computation budget. Further, we propose a light weight dense prediction head that is applicable to both instance edge and mask detection. Finally, we use a penalty reduced focal loss to effectively train the model with point supervision on instance edges, which can reduce annotation costs. We demonstrate highly competitive instance edge detection performance compared to state-of-the-art baselines, and also show that the proposed task and loss are complementary to instance segmentation and object detection.