{
  "paper_id": "f356a9696fdf340c619d59963eb9c6ad",
  "paper_title": "Towards Accurate 3D Person Detection and Localization from RGB-D in Cluttered Environments",
  "paper_authors": [
    "Timm Linder",
    "Dennis Griesser",
    "Narunas Vaskevicius",
    "Kai O. Arras"
  ],
  "publication_date": "",
  "publication_venue": "unknown",
  "abstract": "We focus on the problem of accurately detecting and localizing 3D centroids of persons in RGB-D scenes with frequent heavy occlusions, as often encountered in industrial and service robotics use-cases. While recently, enormous progress has been made in 2D object detection, which is often evaluated in terms of bounding box overlap in image space, robotics systems often rely on metric 3D world coordinates for applications such as human tracking across sensor boundaries, socially aware motion planning or safety and collision avoidance. Starting with a state-of-the-art 2D single-stage detector, we examine how we can robustly lift the coordinates into 3D to outperform the state-of-the-art in RGB-D person detection at 50 frames per second. Evaluation on our Kinect v2 dataset from an intralogistics warehouse indicates that there might be better intermediate representations for this purpose than 2D bounding boxes, such as instance segmentation masks or keypoint estimates. As an alternative strategy, we also compare our method against a recently proposed bottom-up 3D human pose estimation approach. We find that our 2D top-down person detector achieves higher maximum recall, while the bottom-up 3D human pose estimation method can reach higher precision.",
  "keywords": [
    "3D person detection",
    "RGB-D",
    "occlusion",
    "person localization",
    "human pose estimation",
    "intralogistics",
    "instance segmentation",
    "YOLO",
    "depth estimation"
  ],
  "paper_institutions": [],
  "output_language": "zh",
  "provider": "doubao",
  "model": "ark-code-latest",
  "prompt_template": "simple",
  "extracted_at": "2026-08-16T19:47:03.569721Z",
  "template_tag": "simple",
  "summary": "### 摘要\n\n该论文关注于在严重遮挡的RGB-D场景中准确检测和定位人物3D中心点的问题，这在工业和服务机器人应用中常见。尽管2D目标检测取得了巨大进展，但机器人系统通常需要度量3D世界坐标用于人类跟踪、社交感知运动规划以及安全避障。作者从最先进的2D单阶段检测器YOLO v3出发，研究如何鲁棒地将坐标提升到3D，以在50帧每秒的速度下超越RGB-D人物检测的现有水平。在内部物流仓库的Kinect v2数据集上评估表明，相对于2D边界框，实例分割掩码或关键点估计可能是更好的中间表示。作为替代策略，作者还对比了自底向上的3D人体姿态估计方法，发现其2D自上而下检测器能达到更高的最大召回率，而自底向上方法能达到更高的精度。核心机制是利用2D边界框内的深度中值进行简单的3D坐标提升，并进一步集成实例分割掩码以改善遮挡情形下的深度估计。\n\n### 关键要点\n\n- 作者采用基于YOLO v3的2D单阶段检测器作为基线，实现了50Hz的实时处理速度。\n- 通过计算2D边界框内深度值的中值进行朴素3D提升，该方法已经超越了现有的MobilityAids方法的性能。\n- 离线集成实例分割掩码（如Mask R-CNN）可进一步提升AP约2个百分点，实现了更鲁棒的深度估计。\n- 作者录制了新的Kinect v2数据集，包含真实仓库场景和实验室环境，并手动标注了2D边界框和3D中心点轨迹。\n- 对比了自底向上的3D人体姿态估计方法（RGB-D Pose 3D），在0.25m阈值下其能达到更高的最大精度和AP。\n- 作者的方法在0.5m阈值下AP更高，说明2D检测器在高容忍度的定位精度上具有优势。\n- 所有方法在重遮挡、光照变化、低对比度和运动模糊等挑战下均表现不佳。\n- 反射性安全背心可能导致Kinect v2产生完全错误的深度读数，这是实际应用中的限制。\n- 作者建议使用合成遮挡增强技术来提高重遮挡情况下的召回率。\n- 未来的工作可能探索在检测阶段集成深度信息，以改善玻璃和金属表面的反射问题。"
}