Despite significant advances in generic object detection, a persistent performance gap remains for tiny objects compared to normal-scale objects. We demonstrate that tiny objects are highly sensitive to annotation noise, where optimizing strict localization objectives risks noise overfitting. To address this, we propose Tiny Object Localization with Flows (TOLF), a noise-robust localization framework leveraging normalizing flows for flexible error modeling and uncertainty-guided optimization. Our method captures complex, non-Gaussian prediction distributions through flow-based error modeling, enabling robust learning under noisy supervision. An uncertainty-aware gradient modulation mechanism further suppresses learning from high-uncertainty, noise-prone samples, mitigating overfitting while stabilizing training. Extensive experiments across three datasets validate our approach's effectiveness. Especially, TOLF boosts the DINO baseline by 1.2% AP on the AI-TOD dataset.

翻译：尽管通用目标检测领域已取得显著进展，但微小目标与常规尺度目标之间仍存在持续的性能差距。我们证明微小目标对标注噪声高度敏感，优化严格定位目标存在噪声过拟合的风险。为解决此问题，我们提出基于归一化流的噪声鲁棒定位框架TOLF，通过灵活误差建模与不确定性引导优化实现鲁棒学习。该方法通过基于流的误差建模捕获复杂的非高斯预测分布，从而在噪声监督下实现稳健学习。不确定性感知梯度调制机制进一步抑制对高不确定性、易噪声样本的学习，在稳定训练的同时缓解过拟合问题。在三个数据集上的大量实验验证了本方法的有效性。特别地，TOLF在AI-TOD数据集上将DINO基线模型的平均精度提升了1.2%。