Robust Texture Classification Using Local Correlation Features for Thai Buddha Amulet Recognition
In this paper, we propose a development of Thai Buddha amulet identification using simple local correlation features. By using this technique, it has an ability to deal with variety of the amulet materials and colors in the same generation with less computation complexity. Moreover, it is able to apply for semi-controlled environment, which states that the image just has a plain background color that different from the amulet one. This article uses K-nearest neighbors as classification technique. The experiment was done automatically by using amulet images from the internet, which ensured that each image in the same class had a different in light intensity, contrast and color. There were 240 images with 80 classes for training data set and 751 images for test data set. The result shows that the proposed method gains a high recognition rate about 89.35%.