METHOD AND DEVICE FOR CLUSTERING LARGE-SCALE IMAGE DATA

摘要

A method and device for clustering large-scale image data are provided. The method comprises the following steps: a minimum sample number included in a class center of each image data cluster is set, the samples of the image data sample set are divided into class centers of each image data cluster; when the current iteration has been finished, an actual sample number included in the class center of each image data cluster is calculated; said actual sample size included in the class center of each image data cluster is compared with the minimum sample number; the class center of the image data cluster which the actual sample number is less than said minimum sample number is removed, and then the next iteration is performed; the class center of the image data cluster which the actual sample number is equal or greater than said minimum sample number is updated, and the next iteration is performed. The present invention is applied to data cluster in the data searching field, in the condition of performing random selecting an initial center, the class centers which are not representative enough can be removed effectively in the clustering process, and the class centers which are representative enough can be acquired.