PENENTUAN TITIK PERCABANGAN SKELETON MANUSIA DENGAN ATURAN IF-THEN

Abstract

This research proposes If-Then rules to determine the branch point of human skeleton binary images. It has two types, If-Then rules for upper branch (branch of neck, hands and body) and If-Then rules for lower branch (branch of croctch area). If-Then rules contain variation of branch shape on the 24-neighborhood of 5x5 pixel area. This research proposes 12 If-then rules to determine upper branch and 9 If-Then rules to determine lower branch. From 48 skeleton image data, accuracy for determining the branch is 93,75%. Result of this research can be used for determining human pose type from skeleton shape.