启发式采样在训练深度目标探测器中是必要的吗？
Is Heuristic Sampling Necessary in Training Deep Object Detectors?
摘要：
为了在极端的前景-背景不平衡下训练精确的深度目标检测器，启发式采样方法总是必要的，该方法或者对所有训练样本的子集进行重新采样（硬采样方法，\eg偏置采样，OHM），或者使用所有训练样本但对它们进行区分性地重新加权（软采样方法，\eg焦点损失，GHM）。在本文中，我们挑战的必要性，这种硬/软采样方法训练准确的深目标探测器。虽然以前的研究表明，没有启发式采样方法的训练检测器会显着降低准确性，但我们发现这种降低来自不平衡引起的不合理的分类梯度幅度，而不是缺乏重新采样/重新加权。基于我们的发现，我们提出了一个简单而有效的\emph{Sampling-Free}机制，通过初始化和损失缩放来实现合理的分类梯度幅度。与具有多个超参数的启发式采样方法不同，我们的Sampling-Free机制是完全数据诊断的，无需费力的超参数搜索。我们验证了我们的方法在训练基于锚和无锚对象检测器中的有效性，其中我们的方法在COCO和PASCAL VOC数据集上总是比启发式采样方法获得更高的检测精度。我们的无采样机制为解决前景-背景不平衡提供了一个新的视角。我们的代码在\url{https://github.com/ChenJoya/sampling-free}上发布。
英文摘要：
To train accurate deep object detectors under the extreme foreground-background imbalance, heuristic sampling methods are always necessary, which either re-sample a subset of all training samples (hard sampling methods, \eg biased sampling, OHEM), or use all training samples but re-weight them discriminatively (soft sampling methods, \eg Focal Loss, GHM). In this paper, we challenge the necessity of such hard/soft sampling methods for training accurate deep object detectors. While previous studies have shown that training detectors without heuristic sampling methods would significantly degrade accuracy, we reveal that this degradation comes from an unreasonable classification gradient magnitude caused by the imbalance, rather than a lack of re-sampling/re-weighting. Motivated by our discovery, we propose a simple yet effective \emph{Sampling-Free} mechanism to achieve a reasonable classification gradient magnitude by initialization and loss scaling. Unlike heuristic sampling methods with multiple hyperparameters, our Sampling-Free mechanism is fully data diagnostic, without laborious hyperparameters searching. We verify the effectiveness of our method in training anchor-based and anchor-free object detectors, where our method always achieves higher detection accuracy than heuristic sampling methods on COCO and PASCAL VOC datasets. Our Sampling-Free mechanism provides a new perspective to address the foreground-background imbalance. Our code is released at \url{https://github.com/ChenJoya/sampling-free}.
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。
