MEMO：使用掩码边缘预测实现类人的清晰边缘检测
MEMO: Human-like Crisp Edge Detection Using Masked Edge Prediction
提出「掩码边缘预测 + 按置信度渐进定稿」的训练/推理范式，无需专用损失或改网络即可把粗边缘细化到单像素级；其置信度梯度观察与渐进预测策略可迁移到其他密集预测任务（如显著性、抠图、二值分割）的细化环节。
原文依据 we propose a novel progressive prediction strategy that sequentially finalizes edge predictions in order of prediction confidence
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论基于交叉熵损失训练的边缘检测模型往往产生粗厚的边缘预测，与人类通常提供的清晰单像素标注存在偏差。以往实现清晰边缘的方法主要集中于设计专用损失函数或修改网络架构，而我们表明，仅凭精心设计的训练与推理策略就足以达到类人的边缘质量。本文提出掩码边缘预测模型（MEMO），仅使用交叉熵损失即可同时产生准确且清晰的边缘。我们首先构建了一个大规模合成边缘数据集用于预训练 MEMO，以增强其泛化能力。随后在下游数据集上微调时，仅需一个包含 1.2% 额外参数的轻量模块。训练过程中，MEMO 学习在不同输入掩码比例下预测边缘。指导我们推理的一个关键洞见是：粗厚的边缘预测通常呈现置信度梯度，即中心高、向边界逐渐降低。据此，我们提出一种新颖的渐进预测策略，按预测置信度顺序依次确定边缘预测，从而得到更细、更精确的轮廓。我们的方法可获得视觉上令人满意、无需后处理、类人的边缘图，并在清晰度感知评测中优于以往方法。https://github.com/cplusx/MEMO_Edge_Detection
英文原文摘要 1518 字符
Learning-based edge detection models trained with cross-entropy loss often suffer from thick edge predictions, which deviate from the crisp, single-pixel annotations typically provided by humans. While previous approaches to achieving crisp edges have focused on designing specialized loss functions or modifying network architectures, we show that a carefully designed training and inference strategy alone is sufficient to achieve human-like edge quality. In this work, we introduce the Masked Edge Prediction MOdel (MEMO), which produces both accurate and crisp edges using only cross-entropy loss. We first construct a large-scale synthetic edge dataset to pre-train MEMO, enhancing its generalization ability. Subsequent fine-tuning on downstream datasets requires only a lightweight module comprising 1.2% additional parameters. During training, MEMO learns to predict edges under varying ratios of input masking. A key insight guiding our inference is that thick edge predictions typically exhibit a confidence gradient: high in the center and lower toward the boundaries. Leveraging this, we propose a novel progressive prediction strategy that sequentially finalizes edge predictions in order of prediction confidence, resulting in thinner and more precise contours. Our method achieves visually appealing, post-processing-free, human-like edge maps and outperforms prior methods on crispness-aware evaluations. \href https://github.com/cplusx/MEMO_Edge_Detection https://github.com/cplusx/MEMO_Edge_Detection