SMCL：用于长尾识别的显著性掩码对比学习
SMCL: Saliency Masked Contrastive Learning for Long-tailed Recognition
AI总结：
针对长尾数据中背景特征导致预测偏向多数类的问题，提出显著性掩码对比学习方法，通过掩码图像重要部分并在特征空间向少数类移动，有效解耦背景特征相关性，在基准数据集上达到最先进性能。
摘要：
现实世界的数据通常遵循长尾分布，各类别间的样本数量存在高度不平衡。基于不平衡数据进行训练的问题在于，所有类别共有的某些背景特征可能在样本稀缺的类别中未被观察到。因此，这种背景会导致预测偏向“多数”类别。本文提出显著性掩码对比学习（saliency masked contrastive learning），这是一种使用显著性掩码和对比学习来缓解该问题并提升模型泛化能力的新方法。其核心思想是利用显著性检测掩码图像的重要部分，并使用对比学习将掩码后的图像在特征空间中向少数类别移动，使得掩码图像中呈现的背景特征不再与原始类别相关。实验结果表明，该方法在基准长尾数据集上取得了最先进水平的性能。
英文摘要：
Real-world data often follow a long-tailed distribution with a high imbalance in the number of samples between classes. The problem with training from imbalanced data is that some background features, common to all classes, can be unobserved in classes with scarce samples. As a result, this background correlates to biased predictions into ``major" classes. In this paper, we propose saliency masked contrastive learning, a new method that uses saliency masking and contrastive learning to mitigate the problem and improve the generalizability of a model. Our key idea is to mask the important part of an image using saliency detection and use contrastive learning to move the masked image towards minor classes in the feature space, so that background features present in the masked image are no longer correlated with the original class. Experiment results show that our method achieves state-of-the-art level performance on benchmark long-tailed datasets.
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。
