摘要
提出一种基于区域分割的SIFT图像特征提取算法。首先采用改进后的SIFT算法对图像进行计算,同时采用快速分割算法对原始图像进行分割,对照分割后的区域分别取每个区域最显著的几个SIFT特征,最后采用局部降维算法,将高维特征降低到可以接受的低维度特征。实验表明,该方法的运行时间短,对特征提取的位置界定准。
This paper proposes a SIFt image feature extraction algorithm based on region segmentation. Firstly, it introduces an improved SIFT algorithm to extract the features, at the same time, it imposes a fast segmentation algorithm to split the original image into areas, and then takes the most significantly first few SIFT features respectively in each region, at last, it uses a dimensional reduction algorithm to get desire low dimension. The experimentai results show that the operating time of the method is quite short, and it has a great definition of the location of the features.
出处
《信息技术》
2013年第10期126-130,共5页
Information Technology
关键词
自动标注
特征提取
区域分割
SIFT
automatic tagging
feature extraction
region segmentation
SIFT