Abstract
Attribution maps for semantic segmentation are almost always judged by visual plausibility. Yet looking convincing does not guarantee that the highlighted pixels actually drive the model's prediction, nor that attribution credit stays within the target region. These questions require a dedicated evaluation protocol. We introduce a reproducible benchmark that tests intervention-based faithfulness, off-target leakage, perturbation robustness, and runtime on Pascal VOC and SBD across three pretrained backbones. To further demonstrate the benchmark, we propose Dual-Evidence Attribution (DEA), a lightweight correction that fuses gradient evidence with region-level intervention signals through agreement-weighted fusion. DEA increases emphasis where both sources agree and retains causal support when gradient responses are unstable. Across all completed runs, DEA consistently improves deletion-based faithfulness over gradient-only baselines and preserves strong robustness, at the cost of additional compute from intervention passes. The benchmark exposes a faithfulness-stability tradeoff among attribution families that is entirely hidden under visual evaluation, providing a foundation for principled method selection in segmentation explainability. Code is available at this https URL.
Abstract (translated)
语义分割的归因图几乎总是通过视觉合理性来评判。然而，看起来令人信服并不能保证被高亮的像素真正驱动模型的预测，也不能保证归因的贡献度局限于目标区域。这些问题需要专门的评估协议。我们引入了一个可复现的基准，该基准在Pascal VOC和SBD数据集上，跨三种预训练主干网络，测试了基于干预的忠实性、目标外信息泄露、扰动鲁棒性和运行时间。为了进一步展示该基准，我们提出了双证据归因（DEA），这是一种轻量级修正方法，通过基于一致性的加权融合，将梯度证据与区域级干预信号相结合。DEA在两者一致的区域增强强调，并在梯度响应不稳定时保留因果支持。在所有完成的测试中，DEA相较于仅使用梯度的基线，持续提升了基于删除的忠实性，并保持了强鲁棒性，代价是干预过程带来的额外计算开销。该基准揭示了归因方法家族内部的忠实性-稳定性权衡，而这种权衡在视觉评估下完全被隐藏，为分割可解释性中的原则性方法选择提供了基础。代码可在该https URL获取。
URL
https://arxiv.org/abs/2603.22624
