3D DAISY与3D SIFT在YouTube动作数据库上的实验对比
he comparision between 3D DAISY and 3D SIFT descriptor on YouTube
为了更好的获取视频中时间轴上的信息，本文作者在前期工作中提出了3D DAISY描述符。但是由于在前期工作中只是在简单的数据库如KTH和WEIZERMAN验证了该描述的有效性。为了验证3D DAISY描述符在复杂数据库上的表现，本文在动作识别中的常用数据库 YouTube 上将3D DAISY与3D SIFT的识别结果进行对比。通过实验结果验证了3D DAISY 描述符的鲁棒性。
o deal with the information in the additional temporal domain in videos, we proposed a descroptor named 3D DAISY in our previous work. However, this descriptor was only validated on simple dataset, e.g. KTH and Weizerman. In this paper we validate the descriptor on complex dataset YouTube and compare the recognition results between 3D DAISY and 3D SIFT. From the experiment results ,we can find that 3D DAISY is also robust to complex action dataset.
张华、操晓春
计算技术、计算机技术
人工智能动作识别3D DAISY
rtificial IntelligenceAction recognition3D DAISY
张华,操晓春.3D DAISY与3D SIFT在YouTube动作数据库上的实验对比[EB/OL].(2012-12-21)[2025-08-02].http://www.paper.edu.cn/releasepaper/content/201212-556.点此复制
评论