🦷 DeepDent AI: Dental Decay Detection

DeepDent AI is a deep learning-based diagnostic tool designed to assist dentists in identifying potential dental caries (decay) from OPG X-ray images. The project utilizes a Convolutional Neural Network (CNN) built with TensorFlow/Keras and is deployed via a Streamlit web interface.

🚀 Features

• Instant Analysis: Upload an X-ray (JPG/PNG) and get a prediction in seconds.
• High Accuracy: Trained on 5,000+ augmented dental images.
• Confidence Metrics: Displays the AI's certainty level for every diagnosis.
• Binary Classification: Distinguishes between "Healthy" and "Potential Decay."

🛠️ Technical Stack

• Frontend: Streamlit
• Deep Learning: TensorFlow / Keras
• Image Processing: Pillow, NumPy
• Dataset: Dental OPG X-ray Dataset (YOLOv8 format adapted for Binary Classification)

🧠 The Challenge & Solution (Hackathon Highlight)

During development, we encountered a significant Data Bias issue where the model initially achieved 100% accuracy but failed on real-world tests.

The Fix:

1. Data Re-Structuring: Manually separated images into healthy and decay classes based on the presence of YOLO .txt labels.
2. Generalization: Implemented Data Augmentation (rotation, zoom, shear) to prevent the model from memorizing backgrounds.
3. Overfitting Prevention: Added Dropout layers (0.5) and reduced the learning rate to 0.0001 to ensure the model learned actual dental features rather than digital artifacts.

📦 Installation & Usage

1. Clone the repository:

   git clone [https://github.com/your-username/deepdent-ai.git](https://github.com/your-username/deepdent-ai.git)
   cd deepdent-ai