Deep Learning for Cultural Heritage: A Mobile App for Monument Recognition Using Convolutional Neural Networks

Abstract

Convolutional neural networks (CNN) has multi-dimensional features that are inextricably interrelated to the extraordinary identification of visionary objects. Humans have a thirst of eagerness to know about ancient monuments. CNN can be a suitable tool for this purpose, in addition to some image processing approaches. This study has developed a mobile application for monument identification in two classes. The prominent contributions have been performed in the data preprocessing and feature extraction requirement using the Canny approach. In addition, CNN has been applied for the utmost level of recognition of the inputted visionary image to the mobile application using CNN. The proposed mechanism has obtained 90.68% accuracy in the testing period as its utmost accuracy result. The coagulation of CNN and other cutting-edge technology has enhanced the performance of the developed application. Moreover, these features have introduced this implementation as a comparatively desired and required app in daily utilization.