EN
Platform UI is currently English-only. Other languages here track your preference for future translations.
xAgriScan
AI-powered apple leaf disease detection for smarter orchard management
Detects apple leaf diseases with high accuracy using a fine-tuned ResNet-50 model and delivers immediate, actionable agronomic advice via the Gemini Flash API—bridging computer vision and expert-level guidance for precision agriculture.
Try xAgriScan Now🌟 Key Features
Comprehensive disease detection powered by cutting-edge AI
High-Accuracy Detection
Fine-tuned ResNet-50 model identifies 6 apple leaf conditions with precision.
- • Apple Scab (Tavelure)
- • Powdery Mildew (Oïdium)
- • Apple Rust (Rouille)
- • Frogeye Leaf Spot (Tache oculaire)
- • Apple Mosaic Virus (Mosaïque du pommier)
- • Healthy leaves
Instant Expert Advice
AI-powered recommendations through Gemini Flash API provide immediate agronomic guidance after each detection.
Batch Processing
Analyze multiple leaf images simultaneously, streamlining workflow for orchard managers and researchers.
Severity Assessment
Automatic classification of disease severity levels — low, moderate, or high — enabling prioritized intervention.
Modern Web Interface
React-based frontend with intuitive drag-and-drop image upload, built with Vite and Tailwind CSS.
Reliable ML Pipeline
Standard ImageNet preprocessing transforms ensure consistent, production-ready inference with a custom resnet50_xagriscan.pth model.
🚀 Quick Start
Get started with xAgriScan in minutes. Follow these steps to deploy both backend and frontend.
Setup Backend
pip install -r requirements.txt
cp .env.example .env
# Add your GEMINI_API_KEY to .env
python backend/app.pySetup Frontend
cd frontend
npm install
npm run dev💡 The workflow is simple: Upload leaf images → AI Detection via ResNet-50 → Agronomic Advice from Gemini Flash. All preprocessing is handled automatically.
🛠 Tech Stack
Backend
- • Flask API with CORS support
- • PyTorch with fine-tuned ResNet-50
- • Gemini Flash API for agronomic advice
- • PIL for image processing
Frontend
- • React 18 with Vite build tool
- • Tailwind CSS for styling
- • Component-based architecture (UploadZone, ResultCard, BatchResults)
ML Pipeline
- • Custom ResNet-50 model (models/resnet50_xagriscan.pth)
- • Jupyter notebooks for training and analysis
- • Standard ImageNet preprocessing transforms
Model Performance
The model was trained and validated using comprehensive data analysis (see notebooks/data_analysis.ipynb and notebooks/resnet50_finetuning.ipynb). Disease severity is automatically classified as low, moderate, or high based on the detected condition.
🌱 Use Cases
Orchard Management
Early disease detection for preventive treatment, reducing crop loss and optimizing fungicide application.
Agricultural Extension
Support farmers with expert-level diagnostics, even in remote areas with limited access to agronomists.
Research
Disease monitoring and pattern analysis across seasons and regions, enabling data-driven agricultural studies.
Education
Training tool for agricultural students and professionals to learn disease identification and management.
Ready to Transform Your Orchard?
Experience the power of AI-driven apple leaf disease detection. Upload your first image and get instant expert advice.
Launch xAgriScan