🎉 Welcome to StereoScape 🎉

StereoScape is an innovative project that converts 2D images and videos into 3D models using the NeRF (Neural Radiance Fields) model. This web application has a user-friendly frontend built with SvelteKit and TypeScript, and a powerful backend using Python Flask, COLMAP, and TensorFlow.

🚀 Features

• Convert 2D images/videos into 3D models
• User-friendly interface with fluid animations
• High-quality 3D reconstruction using Tiny NeRF model
• Applications in gaming, medical imaging, research, VR/AR, and more

🛠️ Technologies Used

Frontend

• SvelteKit
• TypeScript
• Tailwind CSS
• WebGL animations

Backend

• Python Flask
• COLMAP
• TensorFlow
• MongoDB

⚙️ Installation and Setup

Backend Setup

1. Clone the repository: ```sh git clone https://github.com/LittleZabi/stereo-scape.git cd stereo-scape/backend
2. Create a virtual environment: ```sh python -m venv venv source venv/bin/activate # On Windows, use venv\Scripts\activate
3. Install the required Python libraries: ```sh pip install -r requirements.txt`
4. Start the backend server: ```sh python app.py`\

   Frontend Setup

5. Navigate to the frontend directory: ```sh cd ../web
6. Install the required npm packages: ```sh npm install
7. Start the frontend server: ```sh npm run dev

   📂 Project Structure

• backend/: Contains the Flask application and related Python scripts.
• web/: Contains the SvelteKit frontend application.
• requirements.txt: Lists the Python libraries required for the backend.
• package.json: Lists the npm packages required for the frontend.

📋 Usage

• Upload Images/Videos: Use the user-friendly interface to upload 2D images or videos.
• Processing: The backend processes the images/videos, extracts features using COLMAP, and creates a dataset.
• 3D Reconstruction: The NeRF model processes the dataset to generate high-quality 3D models, viewable from multiple angles.

Project Demo and implementation

Watch the demo of the StereoScape project below:

Click the image above to watch the demo video.

🌟 Contributions

Feel free to contribute to this project by submitting issues or pull requests. Your contributions are highly appreciated!

📄 License This project is licensed under the MIT License.

Made with ❤️ by LittleZabi