RecurSeed 与 EdgePredictMix：面向单阶段和多阶段框架的弱监督语义分割伪标签精炼学习
RecurSeed and EdgePredictMix: Pseudo-Label Refinement Learning for Weakly Supervised Semantic Segmentation across Single- and Multi-Stage Frameworks
AI总结：
针对 WSSS-IL 的漏检与误检问题，提出 RecurSeed 递归精炼伪标签，并结合 EdgePredictMix 边缘感知数据增强，在 PASCAL VOC 2012 和 MS COCO 2014 上达到新最优。
摘要：
尽管仅使用图像级标签的弱监督语义分割（WSSS-IL）具有潜在应用价值，但其较低的性能和实现复杂性仍限制着实际应用。主要原因在于两类现象：(a) 漏检和 (b) 误检：(a) 现有 WSSS-IL 方法精炼得到的类激活图对于大规模物体仍只能表示局部区域；(b) 对于小规模物体，过度激活会使其偏离物体边缘。我们提出 RecurSeed，通过递归迭代交替减少漏检和误检，从而隐式找到使两种误差最小化的最佳结合点。我们还提出一种名为 EdgePredictMix 的新型数据增强（DA）方法，该方法在合并分割结果时利用相邻像素之间的概率差异信息，进一步表达物体边缘，从而弥补现有 DA 方法应用于 WSSS 时的不足。我们在 PASCAL VOC 2012 和 MS COCO 2014 两个基准上均取得了新的最先进性能（VOC val：74.4%，COCO val：46.4%）。代码可在 https://github.com/shjo-april/RecurSeed_and_EdgePredictMix 获取。
英文摘要：
Although weakly supervised semantic segmentation using only image-level labels (WSSS-IL) is potentially useful, its low performance and implementation complexity still limit its application. The main causes are (a) non-detection and (b) false-detection phenomena: (a) The class activation maps refined from existing WSSS-IL methods still only represent partial regions for large-scale objects, and (b) for small-scale objects, over-activation causes them to deviate from the object edges. We propose RecurSeed, which alternately reduces non- and false detections through recursive iterations, thereby implicitly finding an optimal junction that minimizes both errors. We also propose a novel data augmentation (DA) approach called EdgePredictMix, which further expresses an object's edge by utilizing the probability difference information between adjacent pixels in combining the segmentation results, thereby compensating for the shortcomings when applying the existing DA methods to WSSS. We achieved new state-of-the-art performances on both the PASCAL VOC 2012 and MS COCO 2014 benchmarks (VOC val: 74.4%, COCO val: 46.4%). The code is available at https://github.com/shjo-april/RecurSeed_and_EdgePredictMix.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
