Efficiently training accurate deep models for weakly supervised semantic segmentation (WSSS) with image-level labels is challenging and important. Recently, end-to-end WSSS methods have become the focus of research due to their high training efficiency. However, current methods suffer from insufficient extraction of comprehensive semantic information, resulting in low-quality pseudo-labels and sub-optimal solutions for end-to-end WSSS. To this end, we propose a simple and novel Self Correspondence Distillation (SCD) method to refine pseudo-labels without introducing external supervision. Our SCD enables the network to utilize feature correspondence derived from itself as a distillation target, which can enhance the network's feature learning process by complementing semantic information. In addition, to further improve the segmentation accuracy, we design a Variation-aware Refine Module to enhance the local consistency of pseudo-labels by computing pixel-level variation. Finally, we present an efficient end-to-end Transformer-based framework (TSCD) via SCD and Variation-aware Refine Module for the accurate WSSS task. Extensive experiments on the PASCAL VOC 2012 and MS COCO 2014 datasets demonstrate that our method significantly outperforms other state-of-the-art methods. Our code is available at {https://github.com/Rongtao-Xu/RepresentationLearning/tree/main/SCD-AAAI2023}.

翻译：有效训练具有图像级标签的弱监督语义分割（WSSS）的精确深度模型是一项具有挑战性且重要的任务。近年来，端到端WSSS方法因其高训练效率而成为研究焦点。然而，当前方法难以充分提取全面的语义信息，导致伪标签质量低下及端到端WSSS的次优解。为此，我们提出一种简单新颖的自对应蒸馏（SCD）方法，无需引入外部监督即可优化伪标签。我们的SCD使网络能够利用自身导出的特征对应作为蒸馏目标，通过补充语义信息增强网络的特征学习过程。此外，为进一步提升分割精度，我们设计了变分感知精化模块，通过计算像素级变分增强伪标签的局部一致性。最后，基于SCD和变分感知精化模块，我们提出一种高效的端到端Transformer框架（TSCD）用于精确的WSSS任务。在PASCAL VOC 2012和MS COCO 2014数据集上的大量实验表明，我们的方法显著优于其他最先进方法。我们的代码开源于{https://github.com/Rongtao-Xu/RepresentationLearning/tree/main/SCD-AAAI2023}。