🌌 SDSS Galaxy Classification Web App

This project is a Flask-based web application that uses a Convolutional Neural Network (CNN) to classify galaxies from images into 5 categories. It is trained using a dataset from Kaggle and deployed locally for interactive predictions.

---

📁 Project Structure

.
├── static/
│   ├── style.css
│   └── uploads/                       ← (Create this folder manually after training)
├── templates/
│   ├── home.html
│   ├── input.html
│   └── output.html
├── training/
│   ├── Galaxy_Dataset
|       ├── test
|       ├── train
|       └── val
│   ├── Train_dataset
|       ├── Cigar-shaped smooth
|       ├── completely rounf=d smooth
|       ├── edge-on
|       ├── In between smooth
|       └── spiral
│   └── sdss.ipynb                     ← Jupyter notebook to train and save the model
├── SDSSmodel.h5                           ← Saved model after training
├── run.py                             ← Flask web server
├── README.md                          ← This file

---

📦 Dataset Preparation

Download the dataset from Kaggle:
🔗 The Galaxy Classification Dataset

• Ensure your dataset contains exactly 5 classes corresponding to:
  • Cigar-shaped smooth
  • In between smooth
  • Completely round smooth
  • Edge-on
  • Spiral

Organize the dataset in this format:

galaxy_dataset/
├── cigar-shaped smooth/
├── in between smooth/
├── completely round smooth/
├── edge-on/
└── spiral/

---

🧠 Step 1: Train and Save the Model

1. Open sdss.ipynb in Jupyter Notebook or Google Colab.
2. Update the dataset path in the notebook if necessary.
3. Train your model using TensorFlow or Keras.
4. Save the trained model using:

model.save('model.h5')

5. Move the model.h5 file into your project root directory (same as run.py).

---

📂 Step 2: Create Uploads Folder

Create an empty uploads directory inside the static folder to store uploaded images:

mkdir static/uploads

---

▶️ Step 3: Run the App Locally

1. Install dependencies:

pip install flask tensorflow pillow numpy

2. Run the Flask server:

python run.py

3. Open your browser and go to:

http://127.0.0.1:5000/

You’ll be taken to the home page with a “Start Classification” button.

---

📄 HTML Pages

• home.html – Starting page with a navigation button
• input.html – Upload an image
• output.html – View the prediction and confidence score

---

🛠 Tech Stack

• Backend: Python, Flask
• Frontend: HTML, CSS
• ML Framework: TensorFlow / Keras
• Model Format: .h5

---

❗ Notes

• Only image files (.jpg, .jpeg, .png) are supported.
• The model input size used in preprocessing (e.g., (256, 256)) must match what was used during training.
• The app assumes classification into exactly 5 galaxy types.

---

📬 Contact

Maintained by Jayanth Srinivas Bommisetty
For questions or feedback, feel free to raise an issue on GitHub or Gmail: jayanth.b.cse@gmail.com.

---