Color Clustering in the Metal Inscription Images Using ANFIS Filter

Susijanto T. Rasmana (STMIK Surabaya)
Yoyon K. Suprapto (ITS Surabaya)
Ketut E. Purnama (ITS Surabaya)

TELKOMNIKA (Telecommunication Computing Electronics and Control)
Vol 11, No 3: September 2013

Ancient inscriptions are historical records of the past age made on stone or metal media. Currently many ancient inscriptions were damaged because it is too long buried in the ground. This research is the first step to repairing the damaged inscription using Image processing. Efforts to restorations using color clustering with ANFIS method are an early stage to perform letters segmentation in the ancient inscription. The Results of ANFIS clustering method are compared to the spatial fuzzy clustering method (SFCM). The clustering performance measurement is done by measuring root mean square error (RMSE). From RMSE measurements, the average values obtained with ANFIS clustering method is smaller 21.80% than with SFCM. This means there is an increase in clustering performance with ANFIS method compared to SFCM.