Egyptian Hieroglyphics Handwriting Character Classification Using CNN with Explainable AI: SHAP and LIME

Egyptian hieroglyphics, one of the oldest writing and human communication systems. In our modern life, it's challenging to interpret this language due to its complex visual symbols. In this study, we propose a Convolutional Neural Network (CNN) model for classifying Egyptian hieroglyphic handwriting characters. The dataset provides 18 different class of handwritten hieroglyph character images. To enhance model interpretability, we apply explainable AI techniques, specifically SHAP and LIME, to identify regions that influence model predictions. The results that our custom CNN model achieves train accuracy 90.62%, validation accuracy 88.25%, and test accuracy of 84.5%, with specific characters showing high classification performance. Also, AUC score reflects 0.99 to 1.00 for each of the classes.