A novel dataset of Gupta archer type coins for machine learning-based classification

Area of Science:

• Numismatics
• Archaeology
• Computer Science

Background:

• Classifying ancient coins is challenging due to diverse information and cultural heritage.
• Machine learning (ML) advancements require relevant datasets for tasks like coin classification.
• Ancient Gupta archer-type coins are scarce and intricately designed, posing identification difficulties.

Purpose of the Study:

• To present a novel, high-quality dataset of ancient Gupta archer-type coin images.
• To establish a reliable resource for numismatic research and ML applications.
• To facilitate the study of ancient Indian archaeology through coin analysis.

Main Methods:

• Collected coin images from verified private collections and auction houses.
• Ensured all images are of authentic ancient Gupta archer-type coins.
• Annotated coin images using visual examination and numismatic literature.

Main Results:

• A curated dataset of authentic ancient Gupta archer-type coin images has been created.
• The dataset adheres to high standards for numismatic research.
• The dataset provides a foundation for ML-based classification and archaeological insights.

Conclusions:

• The novel dataset addresses the need for reliable resources in numismatics and ML.
• This work supports the identification and study of Gupta archer-type coins.
• The dataset can contribute to a deeper understanding of ancient Indian history and archaeology.