CONTINUOUS LEARNING FOR AUTOMATIC IDENTIFICATION OF OC EO ANTIQUE GLASS JEWELRY

CONTINUOUS LEARNING FOR AUTOMATIC IDENTIFICATION OF OC EO ANTIQUE GLASS JEWELRY

Anh-Khoi Ngo-Ho, Hoang-Bac Bui, Van-Trieu Pham  

Abstract: In terms of the Oc Eo culture, with the diversity found in existing artifacts, we propose using an artificial intelligence system to automatically and thoroughly identify Oc Eo glass jewelry through SEM gemological analysis. We hope that this approach, which has only emerged in archaeology worldwide in the past five years, can improve the quality of identification compared to the classic methods used in Vietnamese archaeology. Our research is driven by the special conditions of the archaeology field, where we aim to utilize evolving continuous learning on archaeological databases, by comparing and selecting the most suitable model that meets the acceptable performance of the dataset. This Recognition Automatic System for Oc Eo Glasses (RAS-OEG), which is capable of distinguishing between Oc Eo and non-Oc Eo glass ornaments, is freely distributed to experts and archaeologists.
Keywords:Antique glass jewelry classification, Oc Eo glass identification, automatic classification of antique glass, Vietnamese glass classification, machine learning in archeology