ojs2 has produced an error Message: WARNING: array_values() expects parameter 1 to be array, integer given In file: /home/journal1/public_html/plugins/blocks/keywordCloud/KeywordCloudBlockPlugin.inc.php At line: 76 Stacktrace: Server info: OS: Linux PHP Version: 5.6.30 Apache Version: N/A DB Driver: mysql DB server version: 10.1.28-MariaDB
ojs2 has produced an error Message: WARNING: array_shift() expects parameter 1 to be array, null given In file: /home/journal1/public_html/plugins/blocks/keywordCloud/KeywordCloudBlockPlugin.inc.php At line: 76 Stacktrace: Server info: OS: Linux PHP Version: 5.6.30 Apache Version: N/A DB Driver: mysql DB server version: 10.1.28-MariaDB
ojs2 has produced an error Message: WARNING: ksort() expects parameter 1 to be array, integer given In file: /home/journal1/public_html/plugins/blocks/keywordCloud/KeywordCloudBlockPlugin.inc.php At line: 79 Stacktrace: Server info: OS: Linux PHP Version: 5.6.30 Apache Version: N/A DB Driver: mysql DB server version: 10.1.28-MariaDB

APLIKASI CLUSTERING CITRA BERDASARKAN CIRI WARNA DAN CIRI TEKSTUR

Fendy Ferdyanciu

Abstract


Nowadays many design applications that apply knowledge of digital image processing, one of which is the application for a group picture or image clustering. Clustering is a way to analyze the data by grouping objects into groups that have similar characteristics between objects. It is difficult to group images if the number of images to much, it was made based on the application of clustering image texture characteristic colors and characteristics, so that more effective and faster in the process of classifying the image. In this application method is used to retrieve Moments Color color characteristics and Haar wavelet method to retrieve texture characteristics. For the determination of initial cluster centroid long sought vector (norm vector), the data which has a difference of the most distant to the average centroid norm that will be the beginning. Furthermore, for the elucidation distance using euclidean distance, then the distance values are compared with the results of the F distribution data grouping, if the distance value is smaller then the data into a single cluster, and if the larger will be separate from the initial group.

The test results demonstrate the success of the grouping is done by the number of cluster 2 at 100%, for the number of cluster 3 at 80%, and for the number of cluster 4 at 40%