融合梯度的协同表示分类改进算法
    点此下载全文
引用本文:张 力,梅伟健.融合梯度的协同表示分类改进算法[J].计算技术与自动化,2017,(4):76-79
摘要点击次数: 234
全文下载次数: 
作者单位
张 力,梅伟健 (陕西师范大学 计算机科学学院陕西,西安 710119) 
中文摘要:协同表示算法是人脸识别中非常典型的基于线性表示的算法,该算法因其操作简单,计算复杂度低等优点已经引起了广泛关注。但是由于协同表示算法直接利用已有的图像二维矩阵进行算法操作,没有考虑图像中像素之间的差异,浪费了一部分图像中的有用特征。在协同表示的基础上,提出了加入融合梯度信息的思想,同时利用水平方向上和垂直方向上的轮廓特征,达到提高算法识别率的结果。实验表明,提出的融合梯度的协同表示算法有效优化了原始的协同表示算法的结果。
中文关键词:人脸识别  协同表示分类  人脸轮廓特征  
 
Improved Collaborative Representation Classification Algorithm Based Fusion Gradient
Abstract:The collaborative representation based classification method is a typical and powerful tool applied in face recognition system.The algorithm has attracted wide attention because of its simple operation and low computational complexity.However,because the collaborative representation algorithm directly uses the existing two-dimensional image matrix,and does not take into account the differences between the pixels in the image,it wastes some of the useful features in the image.On the basis of collaborative representation,this paper puts forward the idea of adding fusion gradient information,and uses the facial contour features in horizontal direction and vertical direction to improve the recognition rate of this algorithm.Experiments show that the proposed algorithm of fusion gradient proposed in this paper can effectively optimize the result of the original collaborative representation algorithm.
keywords:face recognition  collaborative representation based classification method  facial contour features
查看全文   查看/发表评论   下载pdf阅读器