摘要
为更好地实现图像跟踪,寻找更具鲁棒性和计算简便的特征描述子,提出了一种基于核局部不变映射的尺度不变特征转换(scale-invariant feature transform,SIFT)特征描述算法。该算法在继承SIFT算法良好性质的基础上,依据不同空间尺度下能量特征差异性,对尺度内的子图像层数进行细化,以提高稳定特征点的数量。此外,借助核方法的映射特性,解决了局部不变映射法丢失非线性高维特征的问题,形成一种基于核局部不变映射的非线性降维法,进而对特征描述子进行特征重划。实验结果表明,在图像尺度缩放、旋转、模糊、亮度变化等多种场景下,相较现有的主成分分析-SIFT算法,该描述子不但取得更多的稳定特征点,而且计算速度也得到大幅提升。
In order to better realize image tracking, a feature describing scale invariant feature transform （SIFT） algorithm based on the kernel locality preserving proiections （LPP） is proposed to search a more robust and convenient feature descriptor. In the algorithm, SIFT properties are well inherited, and the numbers of sub image plies are refined to increase the amount of stable feature points according to the difference of energy cha- racteristics in every scale space. Moreover, a nonlinear dimension-falling method based on the kernel LPP is formed to redraw feature descriptors By the kernel technique. With this function, the improved I.PP can extract high-dimensional features. The result shows that the novel algorithm can obtain more and better feature points, and its calculating speed quite faster than the principal component analysis SIFT in various scenarios like zoom, rotation, blur and illumination variation of images.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2014年第2期382-389,共8页
Systems Engineering and Electronics
基金
国家自然科学基金(61175029)
国防科技重点实验室基金(9140c610301080c6106)
航空科学基金(20101996009)资助课题
关键词
尺度不变特征转换算法
核局部不变映射
能量特征
核方法
scale invariant feature transform （SIFT） algorithm
kernel locality preserving projection （KLPP）
energy characteristics
kernel technique
