DEtection TRansformer (DETR) and its variants (DETRs) achieved impressive performance in general object detection. However, in crowded pedestrian detection, the performance of DETRs is still unsatisfactory due to the inappropriate sample selection method which results in more false positives. To settle the issue, we propose a simple but effective sample selection method for DETRs, Sample Selection for Crowded Pedestrians (SSCP), which consists of the constraint-guided label assignment scheme (CGLA) and the utilizability-aware focal loss (UAFL). Our core idea is to select learnable samples for DETRs and adaptively regulate the loss weights of samples based on their utilizability. Specifically, in CGLA, we proposed a new cost function to ensure that only learnable positive training samples are retained and the rest are negative training samples. Further, considering the utilizability of samples, we designed UAFL to adaptively assign different loss weights to learnable positive samples depending on their gradient ratio and IoU. Experimental results show that the proposed SSCP effectively improves the baselines without introducing any overhead in inference. Especially, Iter Deformable DETR is improved to 39.7(-2.0)% MR on Crowdhuman and 31.8(-0.4)% MR on Citypersons.

翻译：DEtection TRansformer（DETR）及其变体（DETRs）在通用目标检测中取得了令人瞩目的性能。然而，在密集行人检测中，由于不恰当的样本选择方法导致更多误检，DETRs的性能仍不尽如人意。为解决该问题，我们提出一种简单但有效的DETRs样本选择方法——密集行人样本选择（SSCP），其包含约束引导标签分配方案（CGLA）和可用性感知焦点损失（UAFL）。我们的核心思想是为DETRs选择可学习样本，并根据样本的可用性自适应调节其损失权重。具体而言，在CGLA中，我们提出新的代价函数，确保仅保留可学习的正训练样本，其余均为负训练样本。进一步，考虑样本的可用性，我们设计UAFL根据可学习正样本的梯度比和交并比，自适应分配不同损失权重。实验结果表明，所提SSCP有效提升了基线模型性能，且不引入任何推理开销。特别地，在Crowdhuman上Iter Deformable DETR的MR降至39.7%（-2.0%），在Citypersons上降至31.8%（-0.4%）。