Sri Lankan Institute of Information Technology
An AI-powered pipeline to digitize, restore, and translate ancient Sri Lankan inscriptions from degraded stone surfaces.
Developed by: Lahiru Bandara, Amadhi Hansani, Senod Mesandu, Chamodya Handapangoda
Silent Script introduces a novel deep-learning architecture designed to automate the epigraphical analysis of ancient Sri Lankan inscriptions. By combining image enhancement, instance segmentation, and neural machine translation, the system successfully digitizes heavily degraded stone texts where traditional OCR fails.
Preserving the rich history embedded in ancient stone inscriptions is a race against time. Natural weathering and human interference have left these artifacts severely degraded, making traditional preservation methods insufficient.
Meet the researchers and supervisors behind Silent Script. Feel free to reach out for collaborations or inquiries.
Primary Supervisor
Co-Supervisor
Have a question or want to collaborate? Drop us a message below.
Advanced pre-processing using deep learning to remove noise, correct lighting, and highlight faint carvings on stone.
Generative models reconstruct missing or heavily weathered parts of characters, boosting readability from <20% to>90%.
Instance segmentation algorithms isolate individual lines and words from the continuous, often irregular ancient script.
Precise extraction of individual characters, handling overlapping and joined ancient letters.
A custom Improved U-Net architecture classifies ancient characters with 94.6% accuracy, mapping them to modern digital equivalents.
A specialized Neural Machine Translation model converts the recognized ancient syntax into modern readable text.
Custom convolutional neural network trained on a newly curated dataset of ancient scripts to achieve high-precision classification.
Lahiru BandaraGenerative Adversarial Networks (GANs) specialized in texture filtering and contrast enhancement for lithic surfaces.
Chamodya HandapangodaAn integrated pipeline using spacing detection and Transformer-based NLP models to decode and translate ancient syntax.
Amadhi HansaniInpainting algorithms utilizing contextual awareness to reconstruct eroded or broken characters based on surrounding text.
Senod Mesandu| Model | Precision | Recall | F1-Score | Accuracy |
|---|---|---|---|---|
| Baseline CNN | 0.82 | 0.80 | 0.81 | 81.5% |
| ResNet-50 | 0.89 | 0.88 | 0.88 | 88.2% |
| Improved U-Net (Ours) | 0.96 | 0.94 | 0.95 | 94.6% |
Real-time character recognition and restoration in action.
