Abstract: This paper presents an incremental Laplacian method based on Cam weighted distance. The dimension of the original data are reduced with Laplacian operator, then it obtains the neighbors of each added sample by using the Cam weighted distance and constructs the feature of the current data using such neighbors. The exisiting data embedding results are updated. Experimental results show that the method is effective on data dimension reduction and face expression classification.

Key words: feature extraction, Laplacian operator, Cam weighted distance, data dimension reduction

摘要： 提出一种基于Cam加权距离的增量拉普拉斯方法。对原始数据进行拉普拉斯降维，采用Cam加权距离获得每个添加样本的近邻，由其近邻重构出降维后的插入点，更新近邻发生改变的样本点低维数据。实验结果表明，该方法在数据降维与人脸表情分类方面有较好的效果。

关键词: 特征提取, 拉普拉斯算子, Cam加权距离, 数据降维

CLC Number: 

• TP391.41