Features Guide

This document provides detailed documentation of all SehatScan features, their capabilities, and how to use them effectively.

---

Table of Contents

1. Medical Report Scanning
2. Facial Health Analysis
3. Risk Assessment Generator
4. AI Health Assistant (Chatbot)
5. Analysis History
6. Dashboard Overview
7. User Profile & Settings

---

Medical Report Scanning

Overview

The Medical Report Scanning feature uses AI-powered OCR and analysis to extract and interpret health metrics from medical documents such as lab reports, blood tests, and diagnostic results.

Capabilities

• Document Types Supported:

  • PDF documents (single page)
  • JPEG images
  • PNG images

• File Requirements:

  • Maximum size: 10MB
  • Clear, readable text
  • Standard medical report formats

How It Works

1. Upload: User uploads a medical document via drag-and-drop or file browser
2. Processing: Document is sent to Google Gemini Vision API
3. OCR Extraction: AI extracts all text from the document
4. Intelligent Analysis: AI identifies and structures:
   • Health metrics (values, units, reference ranges)
   • Status classification (normal, low, high, critical)
   • Potential health problems
   • Treatment recommendations
5. Storage: Results are saved to user's analysis history
6. Display: Structured results shown with visual indicators

Output Structure

{
  "raw_text": "Extracted OCR text from document",
  "structured_data": {
    "metrics": [
      {
        "name": "Hemoglobin",
        "value": "14.5",
        "unit": "g/dL",
        "status": "normal",
        "reference_range": "13-17 g/dL"
      },
      {
        "name": "Blood Glucose (Fasting)",
        "value": "126",
        "unit": "mg/dL",
        "status": "high",
        "reference_range": "70-100 mg/dL"
      }
    ],
    "problems_detected": [
      {
        "name": "Elevated Blood Glucose",
        "severity": "moderate",
        "description": "Fasting glucose above normal range"
      }
    ],
    "treatments": [
      {
        "recommendation": "Consult endocrinologist",
        "priority": "high"
      },
      {
        "recommendation": "Monitor diet and carbohydrate intake",
        "priority": "medium"
      }
    ],
    "summary": "Overall health assessment summary"
  }
}

Status Classifications

User Interface

The scan report page includes:

• File Upload Zone: Drag-and-drop area with visual feedback
• File Preview: Image preview before submission
• Progress Indicator: Real-time upload and analysis progress
• Results Display:
  • Metrics table with status badges
  • Problems detected section with severity
  • Treatment recommendations prioritized
  • Collapsible raw OCR text
• Medical Disclaimer: Important notice about professional consultation

Best Practices

1. Image Quality: Upload clear, high-resolution images
2. Orientation: Ensure document is properly oriented
3. Complete Documents: Include all pages of multi-page reports
4. Standard Formats: Common lab report formats work best

---

Facial Health Analysis

Overview

The Facial Health Analysis feature examines uploaded face photos to identify potential health indicators based on visual characteristics such as skin tone, coloration, and visible conditions.

Capabilities

• Supported Formats:

  • JPEG images
  • PNG images

• File Requirements:

  • Maximum size: 10MB
  • Clear face visibility
  • Good lighting conditions
  • Front-facing preferred

Analysis Metrics

Problem Detection Thresholds

Redness Detection:

Yellowness Detection:

Output Structure

{
  "face_detected": true,
  "face_count": 1,
  "visual_metrics": {
    "redness_percentage": 25,
    "yellowness_percentage": 15,
    "skin_tone": "medium"
  },
  "problems_detected": [
    {
      "type": "mild_inflammation",
      "severity": "low",
      "description": "Minor redness detected in facial area",
      "confidence": 0.75
    }
  ],
  "treatments": [
    {
      "recommendation": "Apply gentle moisturizer",
      "priority": "low"
    },
    {
      "recommendation": "Avoid harsh skincare products",
      "priority": "medium"
    }
  ]
}

User Interface

• Photo Upload: Simple drag-and-drop or file selection
• Analysis Preview: View uploaded image before analysis
• Results Display:
  • Visual metrics with percentage bars
  • Detected concerns with severity badges
  • Recommended actions
  • Face detection visualization
• Disclaimer: Note about limitations and professional consultation

Limitations

• Not Diagnostic: Results are indicators, not medical diagnoses
• Lighting Dependent: Poor lighting affects accuracy
• Single Face: Optimized for single face in frame
• Skin Conditions: Limited to visual surface analysis

---

Risk Assessment Generator

Overview

The Risk Assessment Generator combines data from multiple sources to create a comprehensive health risk profile. It integrates medical report data, facial analysis, and user-provided health information.

Data Sources

1. Medical Report Analysis (optional)

   • Lab values and metrics
   • Detected health problems
   • Historical trends

2. Facial Health Analysis (optional)

   • Visual health indicators
   • Skin condition data

3. User-Provided Information (required)

   • Age and gender
   • Current symptoms
   • Medical history
   • Current medications
   • Lifestyle factors

Input Form Fields

Assessment Generation

The AI generates a comprehensive assessment including:

1. Risk Factors Identified

   • Based on combined data analysis
   • Prioritized by severity

2. Health Metrics Summary

   • Key values from reports
   • Trend analysis if multiple reports

3. Recommendations

   • Lifestyle modifications
   • Medical consultations needed
   • Follow-up tests suggested

4. Disclaimer

   • AI limitations acknowledgment
   • Professional consultation advice

Output Format

The assessment is generated in Markdown format for clear presentation:

# Health Risk Assessment

## Patient Profile
- Age: 45 years
- Gender: Male

## Risk Factors Identified

### High Priority
1. **Elevated Blood Glucose**: Value of 126 mg/dL indicates pre-diabetes risk

### Moderate Priority
2. **Mild Facial Inflammation**: Detected redness may indicate stress or skin condition

## Recommendations

1. Schedule appointment with endocrinologist
2. Implement dietary modifications
3. Regular monitoring of blood glucose levels

## Disclaimer
This assessment is AI-generated and should not replace professional medical advice.

PDF Export

Users can export the assessment as a PDF:

1. Click "Download PDF" button
2. Assessment rendered to canvas (html2canvas)
3. Converted to PDF document (jsPDF)
4. Downloaded to user's device

Storage

Generated assessments are saved to the user's analysis history with type risk for future reference and trend tracking.

---

AI Health Assistant

Overview

The AI Health Assistant is a conversational chatbot powered by Google Gemini AI. It provides personalized health guidance based on the user's analysis history and profile.

Capabilities

• Answer Health Questions: General health information and guidance
• Explain Medical Terms: Clarify terminology from reports
• Interpret Results: Help understand analysis results
• Track Trends: Identify patterns across multiple analyses
• Provide Recommendations: Actionable health advice

Context Awareness

The chatbot has access to:

1. User Profile

   • Name and email
   • Account creation date
   • Preferences

2. Analysis History

   • Last 20 analyses
   • All types (face, report, risk)
   • Full structured data

3. Conversation History

   • Last 10 messages
   • Maintains context within session

4. Platform Knowledge

   • SehatScan features and capabilities
   • How to use different tools
   • Limitations and disclaimers

Sample Interactions

User: "What did my last blood test show?"

Assistant: "Based on your most recent blood test from [date], here's a summary:

• Hemoglobin: 14.5 g/dL (normal)
• Blood Glucose: 126 mg/dL (slightly elevated)
• Cholesterol: 195 mg/dL (borderline)

The elevated glucose level is worth monitoring. Would you like me to explain what this might mean?"

---

User: "What is HbA1c and why is it important?"

Assistant: "HbA1c (Hemoglobin A1c) measures your average blood sugar levels over the past 2-3 months. It's important because:

1. Diabetes Indicator: Values above 6.5% suggest diabetes
2. Long-term View: Unlike fasting glucose, it shows patterns over time
3. Treatment Monitoring: Helps track how well diabetes management is working

Based on your history, I don't see an HbA1c test. You might consider asking your doctor about it given your glucose levels."

User Interface

• Chat Window: Full-height conversation display
• Message Input: Text input with send button
• Suggested Questions: Starter prompts for new users
• Message Formatting: Markdown rendering with syntax highlighting
• User Avatar: Profile picture display
• Loading State: Typing indicator during response generation

Safety Features

1. No Diagnosis: Always recommends professional consultation
2. Medical Disclaimers: Clear about AI limitations
3. Fallback Responses: Graceful handling of API errors
4. Scope Limitations: Stays within health guidance boundaries

---

Analysis History

Overview

The History page provides a comprehensive view of all past analyses, enabling users to track their health data over time.

Features

• Chronological View: All analyses sorted by date (newest first)
• Type Filtering: Filter by face, report, or risk analyses
• Search Capability: Find specific analyses
• Pagination: Handle large history efficiently
• Detail View: Click to see full analysis details

Analysis Types

History Card Display

Each history item shows:

• Type Badge: Visual indicator of analysis type
• Date: When the analysis was performed
• Summary: Brief overview of findings
• Key Metrics: Important values or indicators
• Action Button: View full details

Data Retention

• Analyses are stored indefinitely
• Users can request data export
• Deletion available through settings

---

Dashboard Overview

Overview

The main dashboard provides a high-level view of the user's health data and quick access to all features.

Dashboard Components

1. Statistics Cards

   • Total Analyses Count
   • Medical Reports Scanned
   • Face Analyses Performed
   • Risk Assessments Generated

2. Recent Analyses

   • Last 5 analyses
   • Quick access to details
   • Type indicators

3. Quick Actions

   • Scan New Report
   • Analyze Face
   • Generate Risk Assessment
   • Open Chatbot

4. Health Summary

   • Key metrics overview
   • Trend indicators
   • Alerts if any

Navigation

The dashboard sidebar includes:

• Dashboard Home
• Scan Report
• Scan Face
• Risk Assessment
• Chatbot
• History
• Profile
• Settings
• Notifications
• Help

---

User Profile & Settings

Profile Management

Users can manage their profile information:

• Display Name: Update visible name
• Email: View registered email (from Clerk)
• Avatar: Profile picture (managed by Clerk)
• Member Since: Account creation date

Application Settings

• Theme: Light/Dark mode toggle
• Language: Multi-language support (English, Urdu)
• Notifications: Enable/disable notifications
• Data Export: Download all analysis data

Security Settings

• Password: Change password (if using email auth)
• Two-Factor Authentication: Enable 2FA
• Session Management: View active sessions
• Account Deletion: Request account removal

Notification Preferences

• Analysis Complete: Notify when analysis finishes
• Health Alerts: Important health indicators
• Weekly Summary: Periodic health overview
• Email Notifications: Email delivery preferences

---

Feature Comparison Matrix

---

Limitations & Disclaimers

General Limitations

1. Not Medical Devices: SehatScan tools are not medical devices
2. No Diagnosis: Results are informational, not diagnostic
3. Professional Consultation: Always consult healthcare providers
4. AI Accuracy: AI can make errors; verify important findings

Specific Limitations

• OCR Accuracy: Dependent on image quality
• Face Analysis: Surface-level indicators only
• Risk Assessment: Based on available data only
• Chatbot Knowledge: May not have latest medical research

Important Notice

SehatScan is designed to help users understand their health data better. It is not a substitute for professional medical advice, diagnosis, or treatment. Always seek the advice of your physician or other qualified health provider with any questions you may have regarding a medical condition.