本文已被：浏览 2015次 下载 1409次
投稿时间：2013-03-20
投稿时间：2013-03-20
中文摘要: 人脸特征提取是人脸识别过程中的关键步骤.基于局部二值模式(LBP)理论,尝试了一种基于局部三值模式(LTP)纹理特征提取方法.在一个窗口中将邻域像素值与中心像素值和阈值t的和差进行比较,从而根据像素不同位置进行加权求和求出LTP正负值,即可求得灰度值,最终得到LTP特征提取结果.在MATLAB软件平台上进行了算法仿真,实验结果表明,与基于LBP局部纹理描述的算法相比,基于LTP局部纹理描述的算法具有更强的判别能力,特别是对于统一区域的噪声具有更强的鲁棒性.
Abstract:Face feature extraction is very important for face recognition．On the basis of the local two value model (LBP) ，a new local three value model(LTP) method for texture feature extraction is improved．In a window afield of pixel value is compared with the sum and difference between the center pixel valueand the threshold value(t) ．According to the different positionof the pixel the value is in weighted sum and the LTP positive and negative values are obtained，then the gray value and finally the LTP feature extraction results are obtained． The local texture descriptional gorithm(LTP) has more discriminative ability and is more in sensitive to the unification of there gional noise than LBP． And through the givenal gorithm the simulation in software platform and LTPfor image feature extraction are realized．
文章编号：20130323 中图分类号： 文献标志码：
基金项目:
| 作者 | 单位 | |
| 边后琴 | 上海电力学院电子与信息工程学院,上海 200090 | houqinbian@163.com |
| Author Name | Affiliation | |
| BIAN Houqin | School of Electronic and Information Engineering,Shanghai University of Electric Power,Shanghai 200090,China | houqinbian@163.com |
引用文本：
边后琴.人脸图像的LTP特征提取[J].上海电力大学学报,2013,29(3):298-302.
BIAN Houqin.Local Triplet Pattern for Face Image Feature Extraction[J].Journal of Shanghai University of Electric Power,2013,29(3):298-302.
边后琴.人脸图像的LTP特征提取[J].上海电力大学学报,2013,29(3):298-302.
BIAN Houqin.Local Triplet Pattern for Face Image Feature Extraction[J].Journal of Shanghai University of Electric Power,2013,29(3):298-302.