Combination of spatially enhanced bag-of-visual-words model and genuine difference subspace for fake coin detection

摘要: Abstract Fake coins are harmful for society, the detection of which is paramount importance. Due to large quantities fake in real world, it impossible examine them manually. To address this issue, we present an intelligent system automatically detect based on their images. The consists two components: coin image representation and classifier learning. represent image, a new spatially enhanced bag-of-visual-words model, called SEBOVW proposed. Afterwards, improve by building genuine difference subspace. finally represented its projection onto In order discriminate between coins, train using subspace representations. A thorough evaluation proposed has been conducted four datasets, consisting thousands different denominations from countries. Promising experimental results excess 98 % accuracy demonstrate effectiveness validity.