摘要
特征提取是图像识别过程中的重要的一部分,本文利用奇异值分解这种有效的代数特征提取方法,获得图像的有效特征描述,把同一类图像集中的各幅图像的信息矩阵的奇异值向量作为矩阵中的一列构成图像特征矩阵,求出图像类间离散度矩阵和图像类内离散度矩阵,比较它们的相似程度,最终获得有用和有效的特征。
Feature extraction is an important part of the course of image recognition. This paper makes use of the singular value decomposition which is the effective method of algebraic feature extraction to acquire the effective feature description of the image. The method considers the singular vectors of information matrix of each image in the image set as columns of the feature matrix of image set, and calculate the scatter matrix of image within-class and the scatter matrix of image between-class base to find out their similarity. Finally, the useful and effective features are acquired.
出处
《工程数学学报》
CSCD
北大核心
2004年第F12期131-134,共4页
Chinese Journal of Engineering Mathematics