FurcaVision

Home Page
Detection Page (Result Page)

Inspiration# FurcaVision 🦷🧠

Inspiration

Early detection of furcation defects is crucial in preventing advanced periodontal disease and tooth loss. However, identifying these defects in dental X-rays is time-consuming and requires expert analysis. We aimed to create an AI-powered solution that automates this process, making diagnostics faster and more accessible for dental professionals.

What It Does

FurcaVision automatically analyzes dental X-rays to detect furcation defects and bone loss. It:

Classifies images as "With Furcation" or "Without Furcation" using deep learning.
Extracts individual teeth for focused analysis.
Marks both furcation defect areas and bone loss regions for better visualization.

How We Built It

🧠 AI Model

We developed a multi-model deep learning system:

Teeth Extraction Model
Built on YOLOv8, it segments and extracts individual teeth from X-rays for precise analysis.
Furcation Marker Model
Uses U-Net with MobileNetV2 as the backbone to detect and highlight furcation defect areas.
Bone Loss Marker Model
Also based on U-Net with MobileNetV2, it identifies and marks bone loss regions in the X-rays.
Classification Model
A fine-tuned DenseNet201 classifies X-rays into "With Furcation" or "Without Furcation."

⚙️ Tech Stack

Component	Technology
Frontend	React + Chakra UI
Backend	Node.js, Express
Model Serving	TensorFlow
Deployment	Web-hosted platform
Additional	FastAPI, Python

Challenges We Ran Into

Data Limitations: Finding and preprocessing a well-labeled dataset for furcation defects and bone loss was challenging.
Model Optimization: Balancing accuracy while ensuring real-time performance required extensive tuning.
Integration Issues: Coordinating multiple deep learning models to work seamlessly in a web-based system.

Accomplishments That We're Proud Of

✅ Successfully developed a multi-model AI system for furcation detection and bone loss analysis.
✅ Integrated deep learning models into a fully functional web application.
✅ Created a user-friendly interface that simplifies dental diagnostics for professionals.

What We Learned

Implementing multiple AI models requires careful coordination and optimization.
Medical imaging analysis demands high-quality datasets and robust preprocessing.
Efficient model deployment is crucial for a smooth user experience in web-based AI applications.

What's Next for FurcaVision 🚀

Enhanced Accuracy: Further refining models with larger, high-quality datasets.
Additional Dental Diagnostics: Expanding beyond furcation defects to detect other periodontal issues.
User Feedback & Testing: Collaborating with dental professionals to improve usability.
Mobile Compatibility: Extending support for mobile devices for greater accessibility.