Heritage Tourism and Religious Identity: A Deep Neural Network-Based Deductive Classification Approach

This paper is concerned with the classification of Southeast Asian temples with the deep neural network with various images of temples. Two of the most popular Buddhist temples in Thailand are Wat Prakaew and Wat Po. The stunning temples in Thailand draw tens of thousands of visitors every year. Sadly, many tourists to the temple sometimes cannot tell whether their vacation photographs were taken at Wat Prakaew or Wat Po. In this study, we use deep learning and computer vision to identify whether a given photograph was shot at Wat Prakaew or Wat Po. The idea is to use a deep neural network to create a pattern that makes it easy to classify images. The work shows better classification results compared to other machine-learning techniques. The proposed techniques attain 96.33% validation accuracy.