<![CDATA[Feature extraction is an important topic in the classification. Well detected features will be able to increase the level of accuracy for the recognition of an object. Javanese language is a language that is often used by people on the island of Java, Indonesia. Feature extraction of Javanese letter has its own difficulty level for Javanese letter consists of many types of letters and combinations. In this paper, developed an application that can extract a handwritten Javanese letter features.The process is carried out as follows: image that has been previously segmented, skeleton by using the method of Zhang Suen. The results of this skeletonizing will be used to extract image features, which consists of loops, straight lines and curves arch. Loop detection using flood-fill algorithm, line detection using Hough Transform and curve detection using Hough Transform with curve detection algorithm.The input is Javanese letters image that has been segmented previously. The output is the type and number of features that are recognized and can be saved in .csv format. This application is made to the programming language C# with Microsoft Visual Studio 2010 as the IDE.]]>
Copyrights © 2015
