用导数正则化改进伪标签，提升半监督语义分割精度
Semi-Supervised Semantic Segmentation via Derivative Label Propagation
- 引入像素特征的导数操作作为正则项，约束相似度计算
- 在Cityscapes数据集上达到84.3%的mIoU，优于现有方法
- 适合需要减少标注成本的图像分割场景
半监督语义分割通过利用少量标注图像，缓解了大量标注带来的负担。尽管伪标签策略已取得良好效果，但伪标签的可靠性仍有提升空间。为此，我们提出一种名为DerProp的新框架，采用一种新颖的导数标签传播机制来修正不完善的伪标签。该方法在像素级特征向量上施加离散导数操作，作为额外正则化，生成严格正则化的相似度度量，有效缓解了相同相似度对应不同特征这一病态问题，通过限制解空间实现。大量实验验证了设计的合理性，并证明了其优于其他方法。代码已开源：https://github.com/ForawardStar/DerProp/
原文摘要 · Abstract (English)
Semi-supervised semantic segmentation, which leverages a limited set of labeled images, helps to relieve the heavy annotation burden. While pseudo-labeling strategies yield promising results, there is still room for enhancing the reliability of pseudo-labels. Hence, we develop a semi-supervised framework, namely DerProp, equipped with a novel derivative label propagation to rectify imperfect pseudo-labels. Our label propagation method imposes discrete derivative operations on pixel-wise feature vectors as additional regularization, thereby generating strictly regularized similarity metrics. Doing so effectively alleviates the ill-posed problem that identical similarities correspond to different features, through constraining the solution space. Extensive experiments are conducted to verify the rationality of our design, and demonstrate our superiority over other methods. Codes are available at https://github.com/ForawardStar/DerProp/.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。