ayush-medicinal-plant-classifier

AYUSH Medicinal Plant Classifier

🌿 A deep learning model for classifying 7 Indian medicinal plants based on leaf images.

Model Description

This model uses transfer learning with Xception architecture to identify medicinal plants commonly used in AYUSH (Ayurveda, Yoga & Naturopathy, Unani, Siddha, and Homeopathy) medicine.

Classes

The model can identify the following 7 medicinal plants:

1. Aloevera (Aloe barbadensis) - Ghritkumari
2. Amla (Phyllanthus emblica) - Indian Gooseberry
3. Bhrami (Bacopa monnieri) - Water Hyssop
4. Ginger (Zingiber officinale) - Adrak
5. Neem (Azadirachta indica) - Nimba
6. Tulsi (Ocimum sanctum) - Holy Basil
7. Turmeric (Curcuma longa) - Haldi

Model Details

• Base Architecture: Xception (pre-trained on ImageNet)
• Input Size: 299x299x3 RGB images
• Framework: TensorFlow/Keras
• Training Accuracy: 97.56%
• Validation Accuracy: 100%
• Test Accuracy: 98.44%
• Dataset: Indian Medicinal Leaves Image Dataset (719 images)

Usage

Using Hugging Face Inference API

import requests from PIL import Image import io

API_URL = "https://api-inference.huggingface.co/models/Tushansh/ayush-medicinal-plant-classifier headers = {{"Authorization": "Bearer YOUR_HF_TOKEN"}}

def query(image_bytes): response = requests.post(API_URL, headers=headers, data=image_bytes) return response.json() Load your image with open("plant_leaf.jpg", "rb") as f: data = f.read()

Get prediction output = query(data) print(output)

Using Python with transformers

from huggingface_hub import from_pretrained_keras

model = from_pretrained_keras( + REPO_ID + )

Process image and make predictions

Training Details

• Optimizer: Adam
• Loss: Categorical Crossentropy
• Callbacks: EarlyStopping, ReduceLROnPlateau
• Data Augmentation: Rotation, Zoom, Flip, Brightness adjustment

Limitations

• Model is specifically trained for these 7 Indian medicinal plants
• Best results with clear, well-lit leaf images
• May not generalize to other plant species

Ethical Considerations

This model is intended for educational and research purposes. Plant identification should be verified by botanical experts before use in medical applications.

Citation

@misc{{ayush-plant-classifier, author = {{Your Name}}, title = {{AYUSH Medicinal Plant Classifier}}, year = {{2025}}, publisher = {{Hugging Face}} }}

License

Apache 2.0

Contact

For questions or feedback, please open an issue in the repository.