The use of machine learning to classify symbols on cultural monuments to identify their origin and historical period.
Year: 2025
Type: article
Abstract: Identifying and cataloguing cultural heritage objects is a time consuming process that requires expert knowledge. This study explores the application of deep learning models, specifically YOLOv8 and ResNet50, to classify historic buildings by historical epoch and country of origin, respectively. This research was conducted using a dataset of 3,200 images, which featured monuments categorized into four separate historical periods and representing four distinct nations. YOLOv8 detected buildings and classified them into historical epochs while ResNet50 was used for classifying the country of origin. The analysis demonstrated that models achieved a notable degree of effectiveness in identifying both the architectural epochs and the countries of origin.
DOI: 10.35784/jcsi.8020
Authors: Igor Pajura, Sylwester Korga
Institution: Lublin University of Technology
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7117541483