Scopus Indexed Publications

Paper Details


Title
Identification of Construction Era for Indian Subcontinent Ancient and Heritage Buildings by Using Deep Learning
Author
Md. Samaun Hasan, Md. Salah Uddin, Mirza Mohtashim Alam, Mizanur Rahman, Rafita Haque, Rokeya Forhat, Shaikh Muhammad Allayear, S. Rayhan Kabir,
Email
hasan.mct@diu.edu.bd
Abstract

The Indian subcontinent is a south geographic part of Asia continent which consists of India, Bangladesh, Pakistan, Sri Lanka, Bhutan, Nepal, and Maldives. Different rulers or the empire of different periods have built various buildings and structures in these territories like Taj Mahal (Mughal Period), Sixty Dome Mosque (Sultanate Period), etc. From archaeological perspectives, a computational approach is very essential for identifying the construction period of the old or ancient buildings. This paper represents the construction era or period identification approach for Indian subcontinent old heritage buildings by using deep learning. In this study, it has been focused on the constructional features of British (1858–1947), Sultanate (1206–1526), and Mughal (1526–1540, 1555–1857) periods’ old buildings. Four different feature detection methods (Canny Edge Detector, Hough Line Transform, Find Contours, and Harris Corner Detector) have been used for classifying three types of architectural features of old buildings, such as Minaret, Dome and Front. The different periods’ old buildings contain different characteristics of the above-mentioned three architectural features. Finally, a custom Deep Neural Network (DNN) has been developed to apply in Convolutional Neural Network (CNN) for identifying the construction era of above-mentioned old periods.

Keywords
Computational archaeology Digital cultural heritage Computer vision Feature detection Deep neural network (DNN) Convolutional neural network (CNN) Deep learning Machine learning
Journal or Conference Name
International Congress on Information and Communication Technology
Publication Year
2021
Indexing
scopus