人工智能培训

搜索

深度学习论文:一种检测不同域面上地标的两步学习方法(A Two-Step Learning Method For Detecting Landmarks on

[复制链接]
xujinjin1 发表于 2018-9-14 08:54:34 | 显示全部楼层 |阅读模式
xujinjin1 2018-9-14 08:54:34 131 0 显示全部楼层
深度学习论文:一种检测不同域面上地标的两步学习方法(A Two-Step Learning Method For Detecting Landmarks on Faces From  Different Domains)面部基准点的检测显着地受到机器学习领域的快速进步的青睐,特别是在卷积网络中。然而,大多数探测器的准确性在很大程度上取决于大量的注释数据。在这项工作中,我们提出了基于两步学习的adomain自适应方法来检测人类和动物面部的基准点。我们在由不同动物面(猫,狗和马)组成的三种不同数据集上评估我们的方法。实验表明,我们的方法比现有技术表现更好,并且可以使用少量注释数据来利用地标检测来减少对大量注释数据的需求。
The detection of fiducial points on faces has significantly been favored bythe rapid progress in the field of machine learning, in particular in theconvolution networks.However, the accuracy of most of the detectors stronglydepends on an enormous amount of annotated data.In this work, we present adomain adaptation approach based on a two-step learning to detect fiducialpoints on human and animal faces.We evaluate our method on three differentdatasets composed of different animal faces (cats, dogs, and horses).Theexperiments show that our method performs better than state of the art and canuse few annotated data to leverage the detection of landmarks reducing thedemand for large volume of annotated data.深度学习论文:一种检测不同域面上地标的两步学习方法(A Two-Step Learning Method For Detecting Landmarks on Faces From  Different Domains) WOT5Te4R9QO4oyot.jpg
URL地址:https://arxiv.org/abs/1809.04621     ----pdf下载地址:https://arxiv.org/pdf/1809.04621    ----深度学习论文:一种检测不同域面上地标的两步学习方法(A Two-Step Learning Method For Detecting Landmarks on Faces From  Different Domains)
回复

使用道具 举报

您需要登录后才可以回帖 登录 | 立即注册

本版积分规则 返回列表 发新帖

xujinjin1当前离线
新手上路

查看:131 | 回复:0

快速回复 返回顶部 返回列表