【24h】

On a Feature Extraction by LMCUH Algorithm for a Ubiquitous Computing

机译:基于LMCUH算法的普适计算特征提取。

获取原文
获取原文并翻译 | 示例

摘要

This paper proposes an algorithm to detect human faces under various environments. In the first step, information on three color spaces of various features is used to determine histogram of color in the first frame of an image. The histogram obtained by interpolation after combining three color of the image is used as an input of LMCUH network. In the second step, the neural network of Levenberg - Marquadt training algorithm minimizes the error. Next, we find the face in test image by using the trained sets. This method is especially suited for various scales, rotations, lighting levels, or occlusions of the target image. Experimental results show that two - dimensional images of a face can be effectively implemented by using artificial neural network training under various environments. Thus, we can detect the face effectively and this can inevitably lead to the Ubiquitous Computing Environment.
机译:本文提出了一种在各种环境下检测人脸的算法。在第一步中,有关各种特征的三个颜色空间的信息用于确定图像第一帧中的颜色直方图。组合图像的三种颜色后通过插值获得的直方图被用作LMCUH网络的输入。第二步,Levenberg-Marquadt训练算法的神经网络使误差最小。接下来,我们使用训练好的集合在测试图像中找到人脸。该方法特别适合于目标图像的各种比例,旋转,照明级别或遮挡。实验结果表明,通过在各种环境下使用人工神经网络训练,可以有效地实现人脸的二维图像。因此,我们可以有效地检测到人脸,这不可避免地会导致无处不在的计算环境。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号