The ability to detect objects in images at varying scales has played a pivotal role in the design of modern object detectors. Despite considerable progress in removing handcrafted components using transformers, multi-scale feature maps remain a key factor for their empirical success, even with a plain backbone like the Vision Transformer (ViT). In this paper, we show that this reliance on feature pyramids is unnecessary and a transformer-based detector with scale-aware attention enables the plain detector `SimPLR' whose backbone and detection head both operate on single-scale features. The plain architecture allows SimPLR to effectively take advantages of self-supervised learning and scaling approaches with ViTs, yielding strong performance compared to multi-scale counterparts. We demonstrate through our experiments that when scaling to larger backbones, SimPLR indicates better performance than end-to-end detectors (Mask2Former) and plain-backbone detectors (ViTDet), while consistently being faster. The code will be released.

翻译：图像中多尺度目标检测能力在现代检测器设计中至关重要。尽管基于Transformer的检测器在消除手工设计组件方面取得显著进展，多尺度特征图仍是其经验成功的关键因素——即使采用如Vision Transformer (ViT)这类纯骨干网络时也不例外。本文证明这种对特征金字塔的依赖并非必要，通过引入尺度感知注意力机制的Transformer检测器，我们提出纯架构检测器"SimPLR"，其骨干网络与检测头均在单尺度特征上运行。这种纯架构使SimPLR能有效利用自监督学习和ViT扩展策略，在与多尺度检测器的对比中展现出强劲性能。实验表明，当扩展至更大骨干网络时，SimPLR相比端到端检测器（Mask2Former）和纯骨干检测器（ViTDet）均表现出更优性能，同时始终保持更快的推理速度。代码将开源。