DEIMv2融合DINOv3特征，实现从超轻量到大模型的实时检测新标杆。
Real-Time Object Detection Meets DINOv3
- 用DINOv3+空间调优适配器生成多尺度特征，提升细节与语义融合。
- 最大模型仅5030万参数达57.8 AP，最小模型971万参数突破50 AP瓶颈。
- 全系列支持端侧部署，轻量级模型参数减少近半仍保持领先性能。
受简单高效的密集O2O机制启发，DEIM实现更快收敛与更强性能。本文将其扩展至DINOv3特征，推出DEIMv2。该模型涵盖从X到Atto共八种尺寸，覆盖GPU、边缘与移动端部署。针对X、L、M、S型模型，采用DINOv3预训练或蒸馏骨干网络，并引入空间调优适配器（STA），高效将DINOv3单尺度输出转为多尺度特征，融合强语义与细粒度细节以增强检测能力。针对超轻量级模型（Nano、Pico、Femto、Atto），采用HGNetv2并结合深度与宽度剪枝以满足严苛资源约束。结合简化解码器与升级版密集O2O，统一设计使DEIMv2在多种场景下实现卓越性能-成本权衡，刷新当前最优水平。其中，最大模型DEIMv2-X仅需5030万参数即达57.8 AP，超越此前需超6000万参数才能达到56.5 AP的同类模型。小型模型中，DEIMv2-S为首个参数低于1000万（实际971万）且在COCO上突破50 AP的模型，达50.9 AP。即便最轻量级的DEIMv2-Pico（仅150万参数）也实现38.5 AP，参数量仅为YOLOv10-Nano（230万）的50%左右，性能相当。代码与预训练模型已开源。
原文摘要 · Abstract (English)
Driven by the simple and effective Dense O2O, DEIM demonstrates faster convergence and enhanced performance. In this work, we extend it with DINOv3 features, resulting in DEIMv2. DEIMv2 spans eight model sizes from X to Atto, covering GPU, edge, and mobile deployment. For the X, L, M, and S variants, we adopt DINOv3-pretrained or distilled backbones and introduce a Spatial Tuning Adapter (STA), which efficiently converts DINOv3's single-scale output into multi-scale features and complements strong semantics with fine-grained details to enhance detection. For ultra-lightweight models (Nano, Pico, Femto, and Atto), we employ HGNetv2 with depth and width pruning to meet strict resource budgets. Together with a simplified decoder and an upgraded Dense O2O, this unified design enables DEIMv2 to achieve a superior performance-cost trade-off across diverse scenarios, establishing new state-of-the-art results. Notably, our largest model, DEIMv2-X, achieves 57.8 AP with only 50.3 million parameters, surpassing prior X-scale models that require over 60 million parameters for just 56.5 AP. On the compact side, DEIMv2-S is the first sub-10 million model (9.71 million) to exceed the 50 AP milestone on COCO, reaching 50.9 AP. Even the ultra-lightweight DEIMv2-Pico, with just 1.5 million parameters, delivers 38.5 AP, matching YOLOv10-Nano (2.3 million) with around 50 percent fewer parameters. Our code and pre-trained models are available at https://github.com/Intellindust-AI-Lab/DEIMv2
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。