Unlocking Ancient Secrets: A Deep Learning Approach to Cuneiform Symbols Recognition

Cuneiform characters, among the earliest known forms of writing, present significant challenges for automated recognition due to their complex and intricate symbols. This paper introduces a Cuneiform Symbols Recognition System (CSRS) utilizing the VGG16 deep learning model to accurately identify and recognize individual cuneiform symbols within scanned archaeological artifacts. The system specifically focuses on recognizing a sentence from the Code of Hammurabi, characterized by wedge-shaped symbols etched onto clay tablets. The methodology involves data augmentation to expand the dataset, followed by the design and training of the VGG16 model. The model is then tested and validated on the segmented symbols from the Hammurabi cuneiform sentence, achieving high accuracy. This work contributes to the field by demonstrating the effectiveness of transfer learning in deciphering ancient scripts, offering a powerful tool for archaeologists and historians in the study and understanding of ancient civilizations.