🚧 Status: Currently under development. This is a work-in-progress research and engineering project.
DeepDent is a next-generation AI toolkit designed to assist with dental diagnostics using panoramic radiographs (OPG images). It aims to detect pathological signs like tumors and cysts, analyze diagnostic features, and generate structured clinical-style reports — all within a secure, offline desktop application.
- Detect dental tumors or cysts from panoramic X-ray images.
- Predict multiple diagnostic features for annotated regions.
- Generate detailed, printable PDF reports with visual explanations.
- Run securely in a privacy-friendly desktop application.
- 📦 Object Detection: AI model for lesion localization.
- 🧪 Multi-feature Prediction: Custom CNN predicting up to 50 clinical attributes.
- 📊 Explainability: Integrated Grad-CAM, Score-CAM, and pixel attribution maps.
- 🖥️ Desktop App: Windows interface for doctors with full report export.
- 📄 Automated Reports: Structured diagnostic output (with optional language polishing).
Designed for dental clinics, labs, and research use cases where offline, explainable AI support can assist in reviewing panoramic images for secondary opinions or diagnostic aid.
Python,PyTorch,torchvision,TkinterGrad-CAM,Captum,Jinja2,WeasyPrint- Optional:
OpenAI GPT-3.5for natural language polishing
This project is currently in development and not intended for clinical deployment or diagnostic use. All components are research-grade only.
Due to an active legal contract and research collaboration agreement, key architectural, dataset, and clinical implementation details cannot be publicly disclosed until the project is completed and cleared for release. Please refer only to public commits and abstract-level summaries.
To follow our milestone progress, stay tuned to the repository's commits and releases.
For research collaborations or queries, please reach out via [email protected].
Made with 💡 for research and innovation.