You can deploy this model and start earning money today!

baidurjya09/
Plantdiseaseclassifier

Name: Plantdiseaseclassifier
Availability: InStock

Views

No views yet

🌿 Plant Disease Classification System

A state-of-the-art deep learning system for detecting plant diseases with 99.79% accuracy using EfficientNet-B0.

Python

PyTorch

Flask

Accuracy

📊 Model Performance

Metric	Score
Test Accuracy	99.79%
Precision	99.79%
Recall	99.79%
F1-Score	99.79%
Test Samples	9,630 images
Errors	20 out of 9,630

🌾 Supported Crops & Diseases

The model can detect 38 different classes across 14 crop types:

Crops Supported:

🍎 Apple (4 classes)
🍒 Cherry (2 classes)
🍊 Citrus (1 class)
🌽 Corn (4 classes)
🌾 Cotton (4 classes)
🍇 Grape (4 classes)
🍑 Peach (2 classes)
🌶️ Pepper Bell (2 classes)
🥔 Potato (3 classes)
🍚 Rice (3 classes)
🍓 Strawberry (2 classes)
🌱 Sugarcane (1 class)
🍅 Tomato (10 classes)
🌾 Wheat (2 classes)

Total: 38 disease classes (including healthy variants)

📁 Project Structure

3balancedmodel/
├── dataset/                    # 3-class balanced dataset
│   ├── Apple/
│   ├── Cherry/
│   ├── Citrus/
│   ├── Corn/
│   ├── Cotton/
│   ├── Grape/
│   ├── Peach/
│   ├── Pepper_bell/
│   ├── Potato/
│   ├── Rice/
│   ├── Strawberry/
│   ├── Sugarcane/
│   ├── Tomato/
│   └── Wheat/
│
├── model/                      # Trained model files
│   ├── best_model.pt          # EfficientNet-B0 weights
│   └── class_names.json       # Class name mappings
│
├── app/                        # Flask web application
│   ├── app_simple.py          # Main Flask app
│   ├── inference_simple.py    # Inference engine
│   ├── test_api.py            # API testing script
│   ├── templates/             # HTML templates
│   │   └── index.html         # Web interface
│   └── static/                # Static assets
│       ├── css/
│       │   └── style.css      # Styling
│       ├── js/
│       │   └── script.js      # JavaScript
│       ├── uploads/           # Uploaded images
│       └── results/           # Generated results
│
├── results/                    # Evaluation results
│   └── final_evaluation/
│       ├── confusion_matrix.png
│       ├── per_class_metrics.png
│       ├── metrics_comparison.png
│       ├── classification_report.txt
│       ├── per_class_metrics.csv
│       ├── evaluation_summary.json
│       ├── numerical_summary.csv
│       ├── classification_report_table.csv
│       ├── sample_predictions.csv
│       ├── correct_predictions.png
│       └── incorrect_predictions.png
│
├── docs/                       # Documentation
│   ├── API_READY.md
│   ├── INTEGRATION_SUMMARY.md
│   ├── NEW_UI_READY.md
│   ├── HEALTHY_PLANTS_FIXED.md
│   └── PASTE_FEATURE_FIXED.md
│
├── requirements.txt            # Python dependencies
└── README.md                   # This file

🚀 Quick Start

1. Clone the Repository

bash

git clone <your-repo-url>
cd 3balancedmodel

2. Install Dependencies

pip install -r requirements.txt

3. Run the Web Application

bash

cd app
python app_simple.py

The application will start at: http://localhost:5000

🎯 Usage

Web Interface

Open http://localhost:5000 in your browser
Upload an image using one of three methods:
- 🖱️ Click to browse files
- 🤚 Drag & drop images
- 📋 Paste from clipboard (Ctrl+V / Cmd+V)
Click "Analyze Disease"
View results:
- Disease name
- Confidence score
- Description
- Treatment recommendations
- Top 5 predictions

API Usage

Python Example:

python

import requests

url = "http://localhost:5000/predict"
files = {'file': open('plant_image.jpg', 'rb')}
response = requests.post(url, files=files)
result = response.json()

print(f"Disease: {result['predicted_class']}")
print(f"Confidence: {result['confidence']}%")
print(f"Treatment: {result['treatment']}")

cURL Example:

curl -X POST -F "file=@plant_image.jpg" http://localhost:5000/predict

Response Format:

json

{
  "predicted_class": "Apple_Apple_scab",
  "confidence": 89.54,
  "description": "Fungal disease causing dark, scabby lesions on leaves and fruit.",
  "treatment": "Apply fungicides during wet periods. Remove infected plant debris.",
  "top_k_predictions": [
    {"class": "Apple_Apple_scab", "probability": 0.8954},
    {"class": "Corn_Corn_Northern_Leaf_Blight", "probability": 0.0040},
    ...
  ],
  "uploaded_url": "/static/uploads/image.jpg"
}

🏗️ Model Architecture

EfficientNet-B0

Architecture: EfficientNet-B0 (pretrained on ImageNet)
Input Size: 224x224 pixels
Parameters: ~5.3M
Training: Fine-tuned on balanced plant disease dataset
Optimizer: AdamW
Loss Function: CrossEntropyLoss with Label Smoothing (0.1)
Data Augmentation:
- Random rotation (±15°)
- Random horizontal flip
- Color jitter
- Random affine transformations

Training Details

Epochs: 36 (early stopping at epoch 26)
Batch Size: 32
Learning Rate: 1e-4 with cosine annealing
Dataset Split: 70% train, 15% validation, 15% test
Total Images: 76,000 (2,000 per class)
Perfect Balance: All classes equally represented

📊 Dataset Information

Dataset Statistics

Total Images: 76,000
Classes: 38
Images per Class: 2,000 (perfectly balanced)
Image Format: JPG
Resolution: Variable (resized to 224x224 for training)

Data Split

Training: 53,200 images (70%)
Validation: 11,400 images (15%)
Test: 11,400 images (15%)

Class Distribution

All classes have exactly 2,000 images, ensuring:

No class imbalance
Fair representation
Unbiased predictions
High accuracy across all classes

🎨 Web Interface Features

Modern UI

✅ Responsive Design - Works on desktop, tablet, mobile
✅ Drag & Drop - Easy file upload
✅ Clipboard Paste - Copy & paste images directly
✅ Real-time Preview - See images before analysis
✅ Progress Indicators - Loading spinners and animations
✅ Toast Notifications - User-friendly feedback

Visual Feedback

🟢 Green Background - Healthy plants
🔴 Red Background - Diseased plants
✅ Check Icons - Healthy classifications
⚠️ Warning Icons - Disease detections
📊 Confidence Bars - Animated progress bars

Results Display

Disease name with icon
Confidence percentage
Disease description
Treatment recommendations
Top 5 predictions with probabilities
Download report option

📈 Evaluation Results

Overall Performance

99.79% Accuracy on test set
20+ classes with 100% F1-Score
All classes above 98% accuracy
Only 20 errors out of 9,630 predictions

Per-Class Performance

See results/final_evaluation/ for:

Confusion matrices (raw + normalized)
Per-class precision, recall, F1-score
Classification report
Sample predictions (correct + incorrect)
Detailed metrics CSV files

🔧 API Endpoints

Health Check

GET /health

Returns server status and model information.

Predict Disease

POST /predict
Content-Type: multipart/form-data
Body: file (image)

Analyzes uploaded image and returns predictions.

Home Page

GET /

Serves the web interface.

🐳 Docker Deployment (Optional)

Create Dockerfile:

dockerfile

FROM python:3.8-slim

WORKDIR /app

COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

COPY app/ ./app/
COPY model/ ./model/

WORKDIR /app/app

EXPOSE 5000

CMD ["python", "app_simple.py"]

Build and Run:

bash

docker build -t plant-disease-classifier .
docker run -p 5000:5000 plant-disease-classifier

🤗 Hugging Face Deployment

Upload to Hugging Face Hub:

python

from huggingface_hub import HfApi

api = HfApi()

# Upload model
api.upload_file(
    path_or_fileobj="model/best_model.pt",
    path_in_repo="best_model.pt",
    repo_id="your-username/plant-disease-classifier",
    repo_type="model"
)

# Upload dataset
api.upload_folder(
    folder_path="dataset",
    repo_id="your-username/plant-disease-dataset",
    repo_type="dataset"
)

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

Dataset: Curated and balanced plant disease dataset
Model: EfficientNet-B0 architecture
Framework: PyTorch, Flask
UI: Modern responsive web interface

📞 Support

For issues, questions, or contributions:

Open an issue on GitHub
Check documentation in docs/ folder
Review evaluation results in results/ folder

🎯 Key Features Summary

✅ 99.79% Accuracy - State-of-the-art performance
✅ 38 Disease Classes - Comprehensive coverage
✅ 14 Crop Types - Wide variety of plants
✅ Balanced Dataset - 2,000 images per class
✅ Modern Web UI - Beautiful, responsive interface
✅ Three Upload Methods - Click, drag, paste
✅ Real-time Analysis - Results in 1-2 seconds
✅ Treatment Recommendations - Actionable advice
✅ API Access - Easy integration
✅ Complete Documentation - Detailed guides

🚀 Production Ready

This system is production-ready and can be deployed to:

Local servers
Cloud platforms (AWS, GCP, Azure)
Docker containers
Hugging Face Spaces
Heroku, Railway, Render

Built with ❤️ for agricultural innovation and plant health monitoring

Version: 1.0.0
Last Updated: May 2026
Model: EfficientNet-B0
Accuracy: 99.79%

baidurjya09/Plantdiseaseclassifier

🌿 Plant Disease Classification System

📊 Model Performance

🌾 Supported Crops & Diseases

Crops Supported:

📁 Project Structure

🚀 Quick Start

1. Clone the Repository

2. Install Dependencies

3. Run the Web Application

🎯 Usage

Web Interface

API Usage

Python Example:

cURL Example:

Response Format:

🏗️ Model Architecture

EfficientNet-B0

Training Details

📊 Dataset Information

Dataset Statistics

Data Split

Class Distribution

🎨 Web Interface Features

Modern UI

Visual Feedback

Results Display

📈 Evaluation Results

Overall Performance

Per-Class Performance

🔧 API Endpoints

Health Check

Predict Disease

Home Page

🐳 Docker Deployment (Optional)

Create Dockerfile:

Build and Run:

🤗 Hugging Face Deployment

Upload to Hugging Face Hub:

📝 License

🙏 Acknowledgments

📞 Support

🎯 Key Features Summary

🚀 Production Ready

baidurjya09/
Plantdiseaseclassifier