screens\classification_screen library

ClassificationScreen provides the main interface for mosquito species identification and classification.

This screen serves as the core feature of the CulicidaeLab application, enabling users to:

• Image Capture: Take photos using device camera or select from gallery
• AI Classification: Real-time species identification using PyTorch Lite models
• Result Display: Show confidence scores, inference time, and species information
• Disease Information: Access associated disease risks and detailed information
• Observation Submission: Submit findings with location data and notes to the research database

Architecture & State Management

The screen implements the MVVM pattern using:

• ClassificationViewModel for business logic and state management
• Provider pattern for reactive UI updates
• Service locator for dependency injection
• Asynchronous operations for image processing and API calls

User Workflow

1. User selects image source (camera/gallery)
2. Image is captured/selected and displayed
3. AI model processes image and returns prediction
4. Results shown with species info and confidence
5. User can view disease risks or submit observation
6. Optional: Add location data and notes for research

Key Features

• Multi-source Image Input: Camera capture and gallery selection
• Real-time Processing: On-device AI inference with progress indicators
• Rich Results Display: Species information, confidence scores, inference timing
• Disease Risk Assessment: Interactive disease list with detailed information
• Research Integration: Observation submission to CulicidaeLab database
• Error Handling: Graceful handling of classification and network errors
• Responsive Design: Adaptive layouts for different screen sizes

Performance Considerations

• On-device Processing: Uses PyTorch Lite for fast, offline classification
• Image Optimization: Automatic image resizing and preprocessing
• Memory Management: Efficient image handling and model caching
• Background Processing: Non-blocking UI during classification