首个实时端到端多人姿态估计模型，用新解码器和去噪关键点加速训练
DETRPose: Real-Time End-to-End Multi-Person Pose Estimation via Modified Transformer Decoder and Novel Denoising Keypoints
- 改进Transformer解码器实现实时推理，引入去噪关键点提升训练效率
- 参数量减少81%，推理速度提升52%，精度媲美顶尖模型
- 适合需要轻量化实时姿态识别的工业场景应用
多人姿态估计（MPPE）是计算机视觉中的基础任务，旨在检测图像中每个人体关节位置。尽管已有进展，现有基于Transformer的模型仍无法实现实时性能。本文提出DETRPose，首个实时端到端的基于Transformer的2D多人姿态估计模型家族。通过改进GroupPose解码器，实现高效推理；提出新型去噪关键点训练策略，显著加速收敛；扩展变焦损失为关键点相似性变焦损失（Keypoint Similarity VariFocal loss），提升查询质量。大量实验表明，DETRPose在精度上可媲美或超越领先方法，同时仅需五至十倍少的训练轮次。在COCO数据集上，DETRPose-S精度达67.0 AP，与YOLOv8-Pose-X（67.3）和YOLO11-Pose-X（67.2）相当，但参数量仅11.5M（分别为69.4M和58.8M），推理速度更快（2.39ms vs 5.23ms和4.93ms）。在CrowdPose上，DETRPose-X FLOPs降低13.1倍（232.3G vs 3048.1G），精度仅低2%（75.1 vs 76.6 AP），优于ED-Pose-SwinL-5S。在OCHuman上，DETRPose-S超越所有先前模型，展现对分布外数据的鲁棒性。代码已开源。
原文摘要 · Abstract (English)
Multi-person pose estimation (MPPE), which involves detecting body joint positions (keypoints) for every person in an image, is a fundamental task in computer vision. Despite recent advances, no transformer-based model currently achieves real-time performance. This work addresses the latency challenge by introducing DETRPose, the first family of real-time, end-to-end transformer models for multi-person 2D pose estimation. DETRPose significantly enhances the GroupPose decoder, enabling real-time inference. For training, a novel denoising keypoint technique is proposed to accelerate convergence. The varifocal loss is also extended for keypoints, termed Keypoint Similarity VariFocal loss, to improve query quality. Extensive evaluation demonstrates that DETRPose models achieve accuracy comparable to or exceeding that of leading alternatives while requiring five to ten times fewer training epochs. DETRPose-S matches the accuracy of YOLOv8-Pose-X and YOLO11-Pose-X on the COCO dataset (67.0 vs 67.3 and 67.2 in AP) with 81% fewer parameters (11.5M vs 69.4M and 58.8M) and 52% faster inference speed (2.39ms vs 5.23ms and 4.93ms). On the CrowdPose dataset, DETRPose-X has $13.1\times$ fewer FLOPs (232.3G vs 3048.1G) and only $2%$ fewer precision (75.1 vs 76.6 in AP) than ED-Pose-SwinL-5S. On the OCHuman dataset, DETRPose-S surpasses all previous models, showing the robustness of DETRPose on out-of-distribution datasets. Code is available at https://github.com/SebastianJanampa/DETRPose
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。