CVDL-Internship-Project

Vision-Based Diagnosis of Dental and Oral Pathologies

An AI-powered computer vision system for detecting and classifying dental and oral diseases using deep learning and transfer learning models.

Overview

This project uses Computer Vision and Deep Learning to automatically detect multiple dental and oral diseases from images. It compares CNNs trained from scratch with transfer learning and fine-tuned models, achieving up to 95% accuracy using pre-trained architectures.

Diseases Classified

• Cancer Sores
• Cold Sores
• Gum Disease
• Mouth Cancer
• Oral Cancer
• Oral Lichen Planus
• Oral Thrush

Models Used

• CNN (from scratch)
• Transfer Learning
• Fine-tuned pre-trained models:
  • VGG16
  • DenseNet
  • Inception

Tech Stack

• Python
• TensorFlow, Keras
• OpenCV, NumPy, Pandas
• Matplotlib, Seaborn
• Google Colab (GPU)
• Gradio (UI)

Evaluation Metrics

• Accuracy
• Precision, Recall, F1-score
• Confusion Matrix
• ROC & AUC

Results

• Fine-tuned models achieved ~95% accuracy
• Transfer learning significantly outperformed CNN trained from scratch
• Data augmentation improved generalization