YOLOv12: Attention-Centric Real-Time Object Detectors

TL;DR

YOLOv12 introduces an attention-centric architecture that combines the accuracy benefits of attention mechanisms with real-time inference speeds, outperforming previous YOLO versions and other detectors in accuracy and efficiency.

Contribution

This paper presents YOLOv12, a novel attention-based real-time object detector that maintains CNN-like speed while leveraging attention mechanisms for improved accuracy.

Findings

YOLOv12-N achieves 40.6% mAP at 1.64 ms latency on T4 GPU.

YOLOv12 outperforms YOLOv10-N and YOLOv11-N in accuracy.

YOLOv12 surpasses RT-DETR models in speed and efficiency.

Abstract

Enhancing the network architecture of the YOLO framework has been crucial for a long time, but has focused on CNN-based improvements despite the proven superiority of attention mechanisms in modeling capabilities. This is because attention-based models cannot match the speed of CNN-based models. This paper proposes an attention-centric YOLO framework, namely YOLOv12, that matches the speed of previous CNN-based ones while harnessing the performance benefits of attention mechanisms. YOLOv12 surpasses all popular real-time object detectors in accuracy with competitive speed. For example, YOLOv12-N achieves 40.6% mAP with an inference latency of 1.64 ms on a T4 GPU, outperforming advanced YOLOv10-N / YOLOv11-N by 2.1%/1.2% mAP with a comparable speed. This advantage extends to other model scales. YOLOv12 also surpasses end-to-end real-time detectors that improve DETR, such as RT-DETR / RT-DETRv2: YOLOv12-S beats RT-DETR-R18 / RT-DETRv2-R18 while running 42% faster, using only 36% of the computation and 45% of the parameters. More comparisons are shown in Figure 1.