Scholarlink

<p>This paper investigates spurious features in a purpose-built image dataset of Hemudu pottery sherds. To minimize acquisition bias, we constructed a standardized dataset by photographing the fragments in situ at the excavation site, capturing diverse cues such as texture, edges, and shape. Our experiments reveal that fine-grained shape details, though correlated with class labels, function as spurious predictors and undermine robustness. To address this, we propose a supervised contrastive learning method, distinct from deep feature reweighting (DFR). By aligning each sample with its label-preserving augmentations, in which spurious cues are disrupted, we promote augmentation-invariant representations. We further adapt the contrastive loss to more effectively balance positive and negative pairs, and incorporate the flooding strategy to stabilize training. On our Hemudu sherd dataset, the method achieves 97.3% accuracy, outperforming recent baselines and demonstrating its effectiveness in mitigating reliance on spurious shortcuts.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Mitigating spurious features by contrastive learning in pottery sherd recognition

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月4日

查看全文
  • Xin Yu,
  • TianJie Li,
  • ZhenYing Song,
  • HuanDa Lu,
  • YiFan Du,
  • FangYu Wu,
  • Hui Xiao,
  • LiBo Xu

摘要

This paper investigates spurious features in a purpose-built image dataset of Hemudu pottery sherds. To minimize acquisition bias, we constructed a standardized dataset by photographing the fragments in situ at the excavation site, capturing diverse cues such as texture, edges, and shape. Our experiments reveal that fine-grained shape details, though correlated with class labels, function as spurious predictors and undermine robustness. To address this, we propose a supervised contrastive learning method, distinct from deep feature reweighting (DFR). By aligning each sample with its label-preserving augmentations, in which spurious cues are disrupted, we promote augmentation-invariant representations. We further adapt the contrastive loss to more effectively balance positive and negative pairs, and incorporate the flooding strategy to stabilize training. On our Hemudu sherd dataset, the method achieves 97.3% accuracy, outperforming recent baselines and demonstrating its effectiveness in mitigating reliance on spurious shortcuts.