The Classification of Similarities of Buddhist Amulets by the Pattern Recognition Method

Abstract

Nowadays, the Buddhist amulets have been produced many processes such as melting using a metal stamping mold. Each mentioned method effects the different earmark. The earmark is an important indicator of the inspection of Buddhist amulets. This research aims to classify the similarities of Buddhist amulets using the pattern recognition method. The image's front view and back view of Buddhist amulet have been analyzed by three algorithms of Low Discrepancy Sampling, LD, Grayscale Value Pyramid, GV and Laplace-Gradient Wavelet Pyramid, GP. The result was analyzed by using variance analysis at 95% confidence interval with Least-Significant Different, LSD method. It was found that the LD algorithm provided the most accurate and the learning time was 93.732% and 0.071 ms, respectively.