{
  "paper_id": "8cec436321901819470ab0b18aba9de5",
  "paper_title": "Investigation of Customer Behavior Analysis Based on Top-View Depth Camera",
  "paper_authors": [
    "Jumpei Yamamoto",
    "Katsufumi Inoue",
    "Michifumi Yoshioka"
  ],
  "publication_date": "",
  "publication_venue": "2017 IEEE Winter Applications of Computer Vision Workshops (WACVW)",
  "abstract": "Human behavior analysis based on surveillance camera is one of hot topics in security, marketing as well as computer vision and pattern recognition, and these are useful for commercial facilities such as convenience stores or book stores. In general, since surveillance camera is placed on the ceiling near store wall to monitor customer behaviors, the majority of this research utilize human model adapted to a front view of the person because the human shape has high discriminative power for human detection, pose estimation, etc. However, this approach has a problem that customers are often occluded by others in the store. To solve this occlusion problem for behavior analysis, we propose a new human behavior analysis method with top-view depth camera. In this research, for the first step to investigate the effectiveness of the analysis, we suppose the book store situation. Our proposed method is composed two behavior estimators. The first estimator is based on height of hand with depth information and the second is based on SVM with depth and PSA (Pixel State Analysis) based features. The characteristic point of our proposed method is that we utilize these estimators by cascading them. From experimental results with 10 behaviors of 3 subjects, although our research is still exploratory, we have confirmed that our proposed method allows us to obtain 89.5% of average estimation accuracy.",
  "keywords": [
    "customer behavior analysis",
    "top-view depth camera",
    "height of hand",
    "Pixel State Analysis",
    "SVM",
    "book store"
  ],
  "paper_institutions": [
    "Osaka Prefecture University"
  ],
  "output_language": "zh",
  "provider": "doubao",
  "model": "ark-code-latest",
  "prompt_template": "simple",
  "extracted_at": "2026-08-16T19:47:35.801669Z",
  "template_tag": "simple",
  "summary": "### 摘要\n本文针对传统监控摄像头在顾客行为分析中因遮挡导致性能下降的问题，提出一种基于顶视深度相机的顾客行为分析方法，并以书店场景为研究对象。方法采用级联结构：第一步利用手部高度与深度信息估计顾客伸手取书的具体书架层（如顶层、第二层等），第二步利用深度特征与PSA（像素状态分析）特征结合SVM分类器区分其余行为（如平放取书、手持、浏览、观看、左移、右移）。实验使用Kinect v2顶视深度相机，采集3名受试者的10种行为共21,340帧图像，采用三折交叉验证，深度特征维度为64，PSA特征维度为612。结果显示，级联架构将平均估计准确率提升至89.5%，其中伸手类行为改善显著，但“平放”行为准确率仅为48.9%，主要由于归一化后与“浏览”行为轮廓相似。进一步分析表明，不同受试者因个人习惯差异（如头部位置偏移）导致性能波动，受试者A的准确率最低（83.7%），受试者C最高（96.6%）。此外，在连续行为场景中，由于未处理模糊行为，总准确率降至63.5%。未来方向包括设计更具判别力的特征、扩展受试者数量、适应多人和模糊行为，以及应用于便利店等其他场景。\n\n### 关键要点\n- 传统基于前视监控的方法易受顾客相互遮挡影响，本文采用顶视深度相机以缓解遮挡问题并降低隐私风险。\n- 提出的方法分为两级级联：第一级通过深度信息判断手部高度，识别伸手取层的动作；第二级使用深度特征和PSA特征经SVM分类其余行为。\n- PSA特征通过分析像素在时间上的静止/瞬变状态区分背景、静止和运动区域，帮助提取人体轮廓特征。\n- 实验采用Microsoft Kinect v2，安装在书架上方的天花板，分辨率为512×424，共采集3名受试者（身高183、177、172cm）的10种行为，总帧数21,340。\n- 特征维度：深度特征为8×8=64维，PSA特征为612维（基于8×8分块和滑窗统计）。支持向量机使用RBF核，通过LIBSVM实现，采用三折交叉验证（2人训练，1人测试）。\n- 单独使用深度特征、PSA特征及两者组合的平均准确率分别为65.1%、65.0%和71.1%，表明组合特征更有效。\n- 引入级联后平均准确率从71.1%提升至89.5%，尤其显著改善了“顶层”、“第二层”等伸手取书行为。\n- “平放”行为准确率最低（48.9%），常被误判为“浏览”或“第四层”，原因在于归一化后轮廓相似，以及手部被身体遮挡。\n- 不同受试者准确率：A为83.7%，B为88.2%，C为96.6%，差异源于个人习惯（如头位置在归一化图像中偏移），方法对旋转不鲁棒。\n- 连续行为场景下，因未定义模糊的中间行为（如从看架到伸手的过渡），整体准确率仅63.5%，需要进一步处理。\n- 本文方法目前仅适用于单人单次行为，未考虑多人同时出现和未定义行为，未来需扩展至多目标和模糊行为识别。\n- 方法针对书店特定环境设计，需验证在便利店等其他商业场景的泛化性。"
}