🍌 Banana Freshness Checker

Check the freshness of bananas in a snap! This web app uses a Deep learning model to predict the freshness index of a banana based on its image. Simply upload a photo of a banana, and the app will analyze the image and return a freshness index from 0 to 100, where 100 represents maximum freshness.

🌐 Try Here - banana-index.onrender.com

📸 Features

• Banana Freshness Prediction: Upload an image of a banana, and the app will calculate a freshness score.
• Simple, User-Friendly Interface: Easily upload an image and get results with a single click.
• Real-Time Processing: Instantly get predictions without any delay.
• Deployed on Render: Robust and accessible from anywhere.

🖥️ Tech Stack

• Flask: Lightweight web framework for Python, used to build the app's backend.
• PyTorch: Used for the machine learning model and prediction processing.
• Torchvision: Provides access to pre-trained models and utilities for computer vision.
• Render: Cloud hosting platform used to deploy and host the app.

🚀 Getting Started

To run this project locally, follow these steps:

Prerequisites

• Python 3.7 or higher
• pip package installer

Installation

• Clone the repository:

  • git clone https://github.com/RatneshKJaiswal/Banana_Index
  • cd Banana_Index

• Install dependencies:

  • pip install -r requirements.txt

• Download the Model: Ensure that banana_freshness_model.pth (the pre-trained model) is in the project directory. If not available in this repo, download or add it manually.

• Run the App:

  • python app.py

• Access the App: Open your browser and go to http://127.0.0.1:5000 to use the app.

🌈 Usage

• Upload a Banana Image: Choose an image of a banana from your device.

• Get Freshness Index: Click "Check Freshness" to receive a score between 0 and 100, where:

  • 100: Fresh and ripe banana.
  • 0: Spoiled banana.

• Interpreting Results: The freshness index is calculated based on the appearance of the banana, such as color and texture.

📈 Model Details

The app uses a modified GoogLeNet model trained on images of bananas with varying freshness levels. The model analyzes input images and predicts a freshness score based on learned visual patterns.

• Input Size: The model processes images resized to 224x224 pixels.
• Normalization: Images are normalized using the mean and std values typical for pre-trained models.

Detailed Project Description

The Banana Freshness Checker is designed to assist users in determining the ripeness of a banana based on visual appearance. Leveraging deep learning, the app employs a fine-tuned GoogLeNet model trained on banana images with varying stages of ripeness. Users can upload a banana image, and the model predicts a freshness score by analyzing color, texture, and other visual indicators. A score of 100 indicates a perfectly ripe banana, while 0 indicates spoilage. This tool provides an easy and interactive way for users to gauge banana freshness without any subjective interpretation.

🔧 Deployment

This app is deployed on Render. You can also deploy it on other platforms like Heroku or AWS by modifying the deployment settings.

• Deployment on Render

  • Push your code to a GitHub repository.
  • Link the GitHub repository to Render.

• Specify the start command:

  • python app.py

• Set the PORT environment variable in Render settings.

🤝 Contributing

Contributions are welcome! Feel free to submit issues, feature requests, or pull requests.

• Fork the repository.
• Create a new branch.
• Commit your changes.
• Push to your branch and create a pull request.

📄 License

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

📬 Contact

For questions or suggestions, please reach out to ratnesh.kr.jais@gmail.com.

Enjoy using the Banana Freshness Checker! 🍌