摘要翻译 机器勿入

来源：百度知道 编辑：UC知道 时间：2024/09/21 16:25:45
摘要

随着计算机技术和互联网的迅速发展，彩色图像的处理已广泛应用于工业、农业、军事、医学等各个领域。而K-均值聚类方法，是模式识别特别是模式识别中无监督聚类方法的一项非常重要的内容，在图像处理、压缩存储、传输、检索、显示和绘制等方面都有非常重要的应用价值。
然而，K-均值聚类方法的聚类效果受到初始聚类中心选取方法的严重制约，如何能够得到一个好的初始聚类中心，无论是在科学理论领域还是应用领域中对K-均值聚类方法都有比较大的实际意义。所以，我们要对K-均值聚类的初始聚类中心获取进行研究。
现有的种类繁多的K-均值聚类的初始聚类中心选取方法有各自的优缺点。本文重点研究了各方法的优缺点，提出了一种最优的方法——即均方误差最小方式下的聚类（LMS）。最后，我们通过图像以及颜色量化实验，证明了本方法是卓有成效的。
关键词：聚类算法，初始聚类中心，颜色量化

With the computer technology and the rapid development of Internet, color image processing has been widely used in industry, agriculture, military, medical and other fields. The K-means clustering methods, pattern recognition is pattern recognition, especially in the unsupervised clustering method, a very important element, in image processing, compression storage, transmission, retrieval, display and rendering, etc. are very important application value.

However, K-means clustering method the effect of clustering the initial cluster centers selected by the methods of serious constraints, how can we get a good initial cluster centers, whether in the field of scientific theories or applications of K-Means clustering method has more practical significance. Therefore, we want to K-means clustering to obtain the initial cluster center for research.

A wide range of existing K-means clustering method to select the initial cluster centers have their own advantages