{
  "paper_id": "2fc57b48897dbaf2a9985c352fde1dfc",
  "paper_title": "Head Segmentation and Head Orientation in 3D space for Pose Estimation of Multiple People",
  "paper_authors": [
    "Sangho Park",
    "J. K. Aggarwal"
  ],
  "publication_date": "",
  "publication_venue": "4th IEEE Southwest Symposium on Image Analysis and Interpretation",
  "abstract": "We present an algorithm for establishing head orientations of multiple persons in 3D space. Using multiple features from grayscale images (i.e. binary blobs, silhouette contours, and intensity distributions), our algorithm achieves foreground separation, head segmentation, and head-orientation classification, respectively. The information is then combined to form an integrated representation about how the heads of multiple persons are configured in 3D space in order to describe their relative position. The algorithm classifies each head orientation, ranging from 0 to 360 degrees of rotation on a horizontal plane, into eight classes by using a moments-based method. The algorithm can be easily extended to video sequences of image frames for describing how head poses change over time in relation to each person involved in a scene. Experimental results are presented and illustrated.",
  "keywords": [
    "head orientation",
    "head segmentation",
    "stereo vision",
    "multiple people",
    "pose estimation",
    "moments"
  ],
  "paper_institutions": [
    "Computer and Vision Research Center, Department of Electrical and Computer Engineering, The University of Texas at Austin"
  ],
  "output_language": "zh",
  "provider": "doubao",
  "model": "ark-code-latest",
  "prompt_template": "simple",
  "extracted_at": "2026-08-16T19:47:24.201055Z",
  "template_tag": "simple",
  "summary": "### 摘要\n\n本文提出了一种在3D空间中估计多人头部朝向的算法，用于多人姿态估计。该算法针对现有方法多局限于单人、受视角限制、依赖面部特征等不足，采用多特征融合策略：利用二值blob、剪影轮廓和灰度强度分别进行前景分离、头部分割和朝向分类。立体视觉模块通过视差计算多人3D位置，头部分割采用部分椭圆拟合以去除肩部干扰，分类器使用3x4网格的矩特征和K近邻（K=7）方法，将水平面内0到360度旋转划分为8类。实验使用32张训练图像和64张测试图像，结果显示82.8%的正确率，其中15.6%的误差在相邻朝向上（≤45°），仅1.6%为大误差。该方法不依赖面部特征，对低分辨率和像素缺失鲁棒，可扩展到视频序列以分析头部朝向随时间的变化，为理解多人互动中的视觉注意关系提供依据，适用于监控和人机交互等场景。核心机制是利用灰度强度的矩特征进行视图无关的头部朝向分类，并通过立体视觉获得头部3D方位。\n\n### 关键要点\n\n- 算法利用多特征（二值blob、轮廓、强度）分别进行前景提取、头部分割和头部朝向分类。\n- 立体视觉模块采用平行轴相机模型，通过视差计算前景目标的深度，得到多人空间相对位置。\n- 头部分割采用部分椭圆拟合，通过去除底部30度极角区域避免肩膀干扰，优化距离度量。\n- 头部朝向分类器基于视图，不依赖面部特征，使用3x4网格的矩特征，对图像尺寸不敏感。\n- 训练数据包含32张图像（每类4张），测试数据64张（每类8张），朝向分为8类，每类45度间隔。\n- 实验结果显示82.8%的测试图像分类正确，15.6%的分类偏差在相邻朝向（≤45°）内，仅1.6%误差≥45°。\n- 系统可处理低分辨率、缺乏面部特征（眼睛、鼻子等）的头部图像，因矩特征基于强度分布。\n- 适用场景包括监控、人机交互，通过头部朝向推断多人之间的视觉注意关系。\n- 算法可扩展到视频序列，跟踪头部位置和朝向随时间的变化，用于理解交互模式。\n- 约束条件：有效深度范围为2到10米，需要背景建模和阈值设定，对光照变化有一定鲁棒性。"
}