Perio. AI

Perio AI is a mobile-app designed to help dental professionals and patients better understand gum health using AI-powered analysis of 2D dental panoramic X-rays. The platform generates easy-to-understand reports with visual overlays, empowering patients to take proactive steps toward healthier gums. Our mission is to bridge the gap between clinical expertise and patient understanding through accessible technology.

Year
2025

Type
B2C/B2B

The Problem

Busy Dentists

  • Inconsistent Diagnosis – Struggle with variability in how gum disease is detected across clinics, leading to inconsistency and reduced trust.

  • Time Pressure – Busy schedules mean early-stage gum disease is often overlooked during fast-paced exams.

  • Low Confidence in Early Detection – Limited tools to detect gum disease reliably at early stages, which weakens clinical confidence.

  • Patient Communication Burden – Dentists have limited time to explain gum health in detail, and explanations vary, causing inconsistent patient education and reduced engagement.

Confused Patients

  • Unaware of Early Symptoms – Many patients don’t recognize early gum disease, leading to delayed treatment.

  • Confusing Diagnoses – Dental X-rays and clinical language are difficult to understand, making it hard to grasp severity.

  • Lack of Transparency – Limited clarity on treatment urgency, costs, and ability to compare options; second opinions are rare.

  • Treatment Risks – Patients often either delay necessary treatment (worsening condition and cost) or accept procedures without understanding if they’re immediately needed.

OUR SOLUTION

OUR SOLUTION

Perio.AI is an AI-powered diagnostic platform that helps dentists detect gum disease early from dental X-rays.


Unlike traditional dental assessments that rely on subjective interpretation and often miss early-stage cases, Perio.AI applies state-of-the-art deep learning models (YOLO + U-Net) to highlight areas of concern with clear confidence scores and visual explanations. This enables clinicians to catch problems earlier, make more consistent treatment decisions, and build patient trust.

Perio.AI is designed to seamlessly integrate into existing dental workflows while preparing for regulatory pathways, positioning it as a scalable solution for preventive oral healthcare worldwide.

AI HYPOTHESIS

If we train deep learning models (YOLO + U-Net) on large-scale annotated dental X-rays and produce 70%+ confidence level, then the system will detect early-stage gum disease with higher accuracy, consistency, and explainability than human clinicians alone, leading to earlier interventions, improved patient outcomes, and stronger trust in diagnosis.

 Feature Highlights

1.Registration & Onboarding Quiz

The quiz gathers data include the user’s age, smoking status, past gum issues, etc to build a dental health profile that helps with model training as well as give visibility to the user’s dental health over time.

2. X-Ray Upload

Users can upload U.S. standard panoramic dental X-ray for the AI model to generate a heatmap of the epic-center of gum disease and suggest their overall gum health score.

3. User Profile and Resources

Each upload and interpretation will be recorded in the user’s profile to capture their gum health condition overtime and enable early detection.

Based on the user’s preference and profile, the app also recommends educational resources and providers for accessible services.

AI Model Selection

We use a dual-model approach to detect and quantify periodontal disease from X-rays—fast enough for clinic use and precise enough for clinical decisions.

Model Type:

  • YOLOv8: Object detection model used to identify and localize key periodontal indicators (e.g., bone loss areas, periodontal defects) on panoramic dental X-rays.

  • U-Net: Segmentation model used to precisely outline gum and bone structures for more accurate measurement and analysis.

  • Input: 2D panoramic dental X-ray (PNG/JPEG)

  • Output:

    1. Gum health score (0–100)

    2. Highlighted areas of concern (visual overlay)

    3. Text-based summary for patients

AI-Architecture

  • Secure Uploads
    Web application with encrypted transmission for dental X-ray uploads

  • Robust Storage

    AWS S3 for images and DynamoDB for metadata management

  • Data Protection
    Fully isolated VPC with role-based access control for HIPAA compliance