摘要
结合凸轮(Cam)权重距离和正交局部保持投影算法,提出了一种改进的Cam权重正交局部保持投影法。该算法将Cam权重距离代替正交局部保持投影算法在构造邻接图时采用的欧式距离,克服了处理高维数据中数据分布不均匀的问题。运用该改进算法对人脸图像进行特征提取,与局部保持投影算法和正交局部保持投影算法进行实验比较,结果证明,用改进的Cam权重正交局部保持投影法进行人脸识别得到的识别率更高,算法的性能效果更好。
By combining the Cam weighted distance and the orthogonal locality preserving projections algorithm, an improved Cam weighted orthogonal locality preserving projection method was proposed.In this algorithm, the Euclidean distance used to the orthogonal locality preserving projection algorithm is instead of the Cam weight distance, which overcomes the problem of processing the data distribution in high dimensional data.Compared with the local preserving projection algorithm and the orthogonal locality preserving projections algorithm,the results show that the improved Cam weighted orthogonal locality preserving projection method is better in face recognition and recognition rate.
出处
《互联网天地》
2016年第12期109-112,共4页
China Internet
关键词
Cam权重距离
人脸识别
正交局部保持投影算法
特征提取
Cam weighted distance
face recognition
orthogonal locality preserving projections algorithm
feature extraction
