Waste Classification App

A deep learning web app that classifies waste images into Biodegradable or Non-Biodegradable using a custom CNN (Convolutional Neural Network) built with TensorFlow / Keras. Users can either upload a photo or capture one live from their webcam, and get an instant prediction from a friendly Flask web interface.

---

🔗 Live Demo

Deployed URL: https://cnnwasteclassification-1.onrender.com/

---

⚠️ Important Note

• ✅ Local / CLI-based prediction works fine — running python app.py from the terminal gives correct predictions (see example terminal output below).
• ❌ Prediction does NOT work when deployed — the model does not return predictions on the deployed Render server (https://cnnwasteclassification-1.onrender.com/). The app opens and loads the page, but the prediction feature fails after deployment on the server.

Example of a local terminal run (works as expected):

(tf_keras) PS D:\Projects\waste_classification_app> python app.py
WARNING:tensorflow:No training configuration found in the save file, so the model was *not* compiled. Compile it manually.
 * Serving Flask app 'app'
 * Debug mode: on
 * Running on http://127.0.0.1:5000
1/1 [==============================] - 3s 3s/step
127.0.0.1 - - [14/Sep/2026 10:06:46] "POST /predict_webcam HTTP/1.1" 200 -
1/1 [==============================] - 0s 117ms/step
127.0.0.1 - - [14/Sep/2026 10:06:48] "POST /upload HTTP/1.1" 200 -

---

Features

• 📷 Image Upload – predict from any image file (JPG / PNG).
• 🎥 Live Webcam Prediction – capture a frame from your camera and classify it instantly.
• 🧠 CNN Model – powered by a TensorFlow/Keras model trained for waste recognition.
• 🌐 Clean Web UI – simple, mobile-friendly Flask frontend.

---

Tech Stack

Layer	Technology
Backend	Python 3.10, Flask
Machine Learning	TensorFlow 2.12, Keras, NumPy
Image Processing	OpenCV, Pillow
Frontend	HTML, CSS, JavaScript (Webcam API)
Deployment	Render (gunicorn)

---

Folder Structure

waste_classification_app/
│
├── app.py                  # Flask backend: routes, model loading & prediction logic
├── requirements.txt        # Python dependencies
├── .gitignore              # Files/folders excluded from Git
├── .python-version         # Recommended Python version (3.10)
├── README.md               # This file
│
├── model/
│   ├── keras_model.h5      # Trained CNN model (TensorFlow/Keras format)
│   └── labels.txt          # Class labels used during training (Class 1, Class 2)
│
├── templates/
│   └── index.html          # Main page (upload form + webcam capture UI)
│
├── statics/
│   └── style.css           # Stylesheet for the app
│
├── uploads/                # Folder where uploaded images are saved (git-ignored)
│
└── .vscode/
    └── launch.json         # VS Code debug configuration

Note: Local Python environments (venv/, tf_keras/, tf_old/) are git-ignored and should not be committed.

---

Prerequisites

Before getting started, make sure you have installed:

• Python 3.10 – recommended version for TensorFlow 2.12. Download from python.org.
• Git – git-scm.com
• A webcam (optional, only needed for live webcam predictions).

On Windows, make sure "Add Python to PATH" is checked during installation.

---

Getting Started (Clone to Run)

Follow these steps to run the project locally on your machine.

1. Clone the repository

git clone https://github.com/your-username/waste_classification_app.git
cd waste_classification_app

2. Create a virtual environment

Windows:

python -m venv venv
venv\Scripts\activate

macOS / Linux:

python3 -m venv venv
source venv/bin/activate

Your terminal prompt should now show (venv).

3. Install dependencies

pip install --upgrade pip
pip install -r requirements.txt

TensorFlow is a large package, so the installation may take a few minutes.

4. Run the app

python app.py

You should see output similar to:

 * Running on http://127.0.0.1:5000

Open your browser and visit: http://127.0.0.1:5000

---

Usage

1. Upload an image – click Choose File, select an image of waste (e.g., a plastic bottle, paper, food waste), and press Predict.
2. Webcam prediction – allow camera access, point it at an object, and click Capture & Predict.
3. The app will display the result: Biodegradable or Non-Biodegradable.

---

How It Works

1. Model loading – The app loads the pre-trained CNN (model/keras_model.h5) once at startup (app.py).
2. Preprocessing – Every input image is resized to 224 × 224 pixels and normalized to a 0–1 range.
3. Prediction – The model outputs a probability score for each class:
   • Class 1 → Biodegradable
   • Class 2 → Non-Biodegradable
4. Response – The result is rendered back on the page via Flask templates.

The model is compatible with a Google Teachable Machine exported Keras model (it maps directly to the same input shape and class count).

---

Deployment (Render)

This project is ready to deploy on Render:

1. Push the repository to GitHub.

2. In Render, create a New Web Service and connect the repo.

3. The build will install dependencies from requirements.txt and start with gunicorn (included in requirements).

4. Start command (set in Render if prompted):

   gunicorn app:app

Set the Python runtime to 3.10 in the Render service settings for best TensorFlow compatibility.

---

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

---

License

This project is for educational purposes. If you plan to use it commercially, make sure the model and dataset used for training are properly licensed.