Pembuatan Aplikasi Segmentasi Dan Klasifikasi Jerawat Dengan Metode Region Growing Dan Self Organizing Map

Abstract

Acne vulgaris is a common skin disease found in men of all ages and sexes. Acnes have different types according to their severity.

In this research, an application were developed to detect, do segmentation, and process the classification an acne object in human’s face. The process begins with the insertion of several seed points on a photo. Each of those seed points were developed further into a region that mask the whole acne using Region Growing Method. Afterward, the regions were grouped together with other acne of similiar features using Self Organizing Map.

According to the experiment’s result, the Region Growing Method gives a satisfying result to do a segmentation on an acne object. But it should be pointed out that every different acne object requires different threshold to achieve an ideal result. Self Organizing Map gives an undesirable result, as the input photos with different skin colors and lightning conditions affect the accuraccy of the result.