RankED: Addressing Imbalance and Uncertainty in Edge Detection Using Ranking-based Losses

RankED proposes a unified ranking-based approach to address the imbalance and uncertainty problems in edge detection, outperforming previous studies on various datasets.