Image-based monument classification using bag-of-word architecture

This paper presents classification of Archaeological monument images by taking into consideration various visual feature representations such as shape, texture and grouping of features into words. There are a number of architectures present for archaeological monuments in India. The purpose of recognition and classification of these monuments has been carried out by taking into consideration the visual features of these architectures. There are a number of representations for shape and texture. It has been seen that the Bag-of-words architecture is found out to be better than other techniques for the purpose of representation and classification. This is seen to have achieved a result of 96.07% accuracy using SVM classifier. It is better than the other hand crafted features that have been used.