Edge-Boundary-Texture Loss: A Tri-Class Generalization of Weighted Binary Cross-Entropy for Enhanced Edge Detection

TL;DR

This paper introduces the EBT loss, a tri-class approach for edge detection that improves accuracy by explicitly modeling edges, boundaries, and textures, outperforming traditional binary loss methods.

Contribution

The paper presents the EBT loss, a novel tri-class loss function that generalizes WBCE and enhances edge detection performance with minimal hyperparameter tuning.

Findings

EBT loss outperforms existing methods on multiple benchmarks.

The loss generalizes binary cross-entropy, with WBCE as a special case.

Model training with EBT is robust to hyperparameter variations.

Abstract

Edge detection (ED) remains a fundamental task in computer vision, yet its performance is often hindered by the ambiguous nature of non-edge pixels near object boundaries. The widely adopted Weighted Binary Cross-Entropy (WBCE) loss treats all non-edge pixels uniformly, overlooking the structural nuances around edges and often resulting in blurred predictions. In this paper, we propose the Edge-Boundary-Texture (EBT) loss, a novel objective that explicitly divides pixels into three categories, edge, boundary, and texture, and assigns each a distinct supervisory weight. This tri-class formulation enables more structured learning by guiding the model to focus on both edge precision and contextual boundary localization. We theoretically show that the EBT loss generalizes the WBCE loss, with the latter becoming a limit case. Extensive experiments across multiple benchmarks demonstrate the superiority of the EBT loss both quantitatively and perceptually. Furthermore, the consistent use of unified hyperparameters across all models and datasets, along with robustness to their moderate variations, indicates that the EBT loss requires minimal fine-tuning and is easily deployable in practice.