Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Advanced Deep Learning Approaches for Automated Recognition of Cuneiform Symbols

Shahad Elshehaby, Alavikunhu Panthakkan, Hussain Al-Ahmad, Mina Al-Saad

TL;DR

The study tackles automated recognition and translation of cuneiform symbols using five deep-learning models trained on a 14,100-sample dataset, with evaluation on Hammurabi Law 1 to extract Akkadian meanings and English translations. It demonstrates that EfficientNetV2M achieves 98.31% accuracy on Hammurabi Law 1, while VGG16 reaches 88.87%, highlighting strong model performance and practical translation capability. A linguistic analysis of Akkadian-Arabic links situates the work within historical language study and supports translation validation. The authors propose ensemble and hybrid architectures for further improvements and emphasize the potential of deep learning to advance archaeology and the preservation of ancient manuscripts.

Abstract

This paper presents a thoroughly automated method for identifying and interpreting cuneiform characters via advanced deep-learning algorithms. Five distinct deep-learning models were trained on a comprehensive dataset of cuneiform characters and evaluated according to critical performance metrics, including accuracy and precision. Two models demonstrated outstanding performance and were subsequently assessed using cuneiform symbols from the Hammurabi law acquisition, notably Hammurabi Law 1. Each model effectively recognized the relevant Akkadian meanings of the symbols and delivered precise English translations. Future work will investigate ensemble and stacking approaches to optimize performance, utilizing hybrid architectures to improve detection accuracy and reliability. This research explores the linguistic relationships between Akkadian, an ancient Mesopotamian language, and Arabic, emphasizing their historical and cultural linkages. This study demonstrates the capability of deep learning to decipher ancient scripts by merging computational linguistics with archaeology, therefore providing significant insights for the comprehension and conservation of human history.

Advanced Deep Learning Approaches for Automated Recognition of Cuneiform Symbols

TL;DR

The study tackles automated recognition and translation of cuneiform symbols using five deep-learning models trained on a 14,100-sample dataset, with evaluation on Hammurabi Law 1 to extract Akkadian meanings and English translations. It demonstrates that EfficientNetV2M achieves 98.31% accuracy on Hammurabi Law 1, while VGG16 reaches 88.87%, highlighting strong model performance and practical translation capability. A linguistic analysis of Akkadian-Arabic links situates the work within historical language study and supports translation validation. The authors propose ensemble and hybrid architectures for further improvements and emphasize the potential of deep learning to advance archaeology and the preservation of ancient manuscripts.

Abstract

This paper presents a thoroughly automated method for identifying and interpreting cuneiform characters via advanced deep-learning algorithms. Five distinct deep-learning models were trained on a comprehensive dataset of cuneiform characters and evaluated according to critical performance metrics, including accuracy and precision. Two models demonstrated outstanding performance and were subsequently assessed using cuneiform symbols from the Hammurabi law acquisition, notably Hammurabi Law 1. Each model effectively recognized the relevant Akkadian meanings of the symbols and delivered precise English translations. Future work will investigate ensemble and stacking approaches to optimize performance, utilizing hybrid architectures to improve detection accuracy and reliability. This research explores the linguistic relationships between Akkadian, an ancient Mesopotamian language, and Arabic, emphasizing their historical and cultural linkages. This study demonstrates the capability of deep learning to decipher ancient scripts by merging computational linguistics with archaeology, therefore providing significant insights for the comprehension and conservation of human history.
Paper Structure (14 sections, 4 equations, 5 figures, 3 tables)

This paper contains 14 sections, 4 equations, 5 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Dataset
  3. Methodology
  4. Model Design and Architecture
  5. Model Training and Optimization
  6. Performance Assessment and Metric Comparison
  7. Comparative Linguistic Investigation
  8. Results and Discussion
  9. Preprocessing and Character Segmentation
  10. Model-Based Character Recognition
  11. Evaluation and Analysis
  12. Cuneiform Character Translation and Linguistic Insights
  13. Observations
  14. Conclusion

Figures (5)

  • Figure 1: Flowchart of the proposed Cuneiform Recognition methodology
  • Figure 2: A log-scale comparison of loss values across deep learning models
  • Figure 3: Scanned image of Hammurabi's Law 1 b10
  • Figure 4: Prediction results of the best-two performing deep learning models compared with the ground truth data
  • Figure 5: Sample of the recognized cuneiform words